Table of Content

25 August 2024, Volume 33 Issue 8

Previous Issue   

Theory Analysis and Methodology Study

Agile Earth Observation Satellite Scheduling with Time-dependent Profits

PENG Guansheng, SONG Guopeng, LIU Xiaolu, HE Yongming, XING Lining

2024, 33(8):  1-7.  DOI: 10.12005/orms.2024.0243

Asbtract ( )   PDF (1135KB) ( )  

References | Related Articles | Metrics

Earth Observation Satellites (EOS) usually refer to a type of satellite platform equipped with optical remote sensors to obtain optical images of the earth’s surface to respond to the needs of different users. EOS are widely used in fields such as weather forecasting, disaster monitoring, natural resource exploration, and military reconnaissance due to their advantages such as wide coverage, long observation duration, and having no airspace and national boundaries. How to make effective use of limited satellite resources and improve the efficiency of reconnaissance satellite task scheduling has become a key problem to be solved urgently. The scheduling problem of EOS is given a set of observation targets with different benefits, and under the condition of satisfying a series of satellite operation and resource constraints, to select and rank a part of the observation targets, formulate a reasonable observation scheduling plan, and achieve the maximum observation benefit change. Due to the influence of look angles on the image quality, the time-dependent profits should be considered in the problem, which undoubtedly increases its difficulty. According to different attitude maneuvering capabilities and working mechanisms, EOS can be divided into two types: traditional Non-agile EOS and Agile EOS (AEOS). AEOS have maneuverability in three axes of roll, pitch and yaw. Among them, the yaw angle refers to the angle between the rotation of the satellite around the normal vector of the ground plane and the running track. Pitching maneuverability enables the satellite to conduct observation activities before or after passing directly above the target, that is, the target is visible to the satellite within a period of time with the time passing through the nadir point as the midpoint, and this period of time is called visible time window. The yaw maneuvering capability does not need the satellite to take images along the direction of the running track. Therefore, the task scheduling of non-sensitive satellites only needs to consider which targets are selected for observation, while the task scheduling of agile satellites also needs to determine the observation time of the targets within the visible time window. Obviously, the flexible attitude maneuvering ability can greatly improve the working efficiency of agile satellites, but at the same time, the increase in solution space and difficulty also poses huge challenges to the design of satellite task scheduling algorithms. Moreover, time-dependent characteristic increases the difficulty. This characteristic originates from the application scenario of satellite image acquisition: generally speaking, images taken from different observation angles ( different observation times) will result in different image quality. For example, at the nadir point, the straight-line distance between the satellite and the observation target is the shortest, and the image distortion caused by the observation angle is small when shooting images, so the quality of the captured image is the best; on the contrary, at the position away from the nadir point, the distance is the farthest, and the larger the viewing angle, the worse the image quality.
   Based on these problem characteristics, we build an integer programming model. To efficiently solve this model, we propose a branch-and-price exact algorithm and an efficient primal heuristic with the guarantee of the solution quality. The proposed exact algorithm is the first to solve the multi-orbit agile satellite scheduling to optimality. The original formulation is decomposed into a master problem and several identical pricing problems. Each pricing problem is equivalent to a single-orbit scheduling problem, which corresponds to a well-known resource constrained elementary shortest path problem. The pricing problem is solved by a bidirectional dynamic programming algorithm with Decremental Space State Relaxation technique. The basic idea is to firstly assign a forward label to the virtual initial node and a backward label to the virtual end node, and gradually extend the label to the rest of the reachable nodes from the forward and backward directions through the defined expansion function. A new label is generated on the new node, and according to the extended termination condition, the forward label and backward label are spliced together to form a complete path, and the path with the largest weight is taken as the optimal solution to the pricing sub-problem. The primal heuristic algorithm provides high-quality initial solutions for branch and price algorithm, and the optimality gap of the heuristic solutions can be given. The idea of the primal heuristic is to use the column generation algorithm to solve the master problem on the root node, according to the information of the fractional solution, fix the single-orbit scheduling scheme with value equals to 1, and continue to solve the master problem with the remaining orbits with value less than 1, until the schemes during all orbits are determined.
   We generate a large number of benchmark instances and conduct experiments in this research. The observation target in the calculation example is uniformly and randomly generated from a rectangular area according to the latitude and longitude. The range of the rectangular area is latitude 3°N-53°N and longitude 74°E-133°E. The results show that the branch and price algorithm can find the optimal integer feasible solution in most cases with only a few dozen steps of branch search. It can solve the 150-target instances with no more than 500 seconds on average. With an increase in the scale of the calculation example, the integrity gap of the calculation instance becomes larger, and the number of branch nodes increases. The branch pricing algorithm needs to spend more computing time getting the optimal solution. Furthermore, the proposed primal heuristic outperforms the state-of-the-art heuristic in terms of solution quality. For the 150-target instances, its optimality gap is no more than 0.3% on average.


Incentive Contract Design for Complex Weapon Equipment Delivery Based on Fairness Preference

ZHANG Yehui, ZHU Jianjun, ZHANG Hang

2024, 33(8):  8-14.  DOI: 10.12005/orms.2024.0244

Asbtract ( )   PDF (1023KB) ( )  

References | Related Articles | Metrics

With the changes in the international situation and the development of the strategic game pattern of major powers, the modernization process of the country’s national defense is accelerating, and new equipment is being added to the army at an accelerated pace, which has brought new challenges to the complex weapon equipment manufacturing enterprise. Improving the delivery efficiency of complex weapon equipment has become a hot issue of concern in all circles. In order to scientifically build the complex weapon equipment delivery system, the key factor that restricts the improvement of delivery efficiency, namely the efficiency of dealing with delivery problems, is studied. Due to the particularity of the complex weapon equipment, it needs to be rechecked during the delivery stage, leading to a series of delivery problems. The continuity of the causes of delivery problems leads to the buck-passing and low enthusiasm for problem handling between departments. The poor quality of problem handling leads to rework of problems. And then, the long problem-handling cycle will affect delivery efficiency and customer satisfaction. According to the data, the average single problem-handling cycle of a certain type of weapon equipment is 3 days, and the number of delivery problems of this type reaches hundreds. Compared with the seven-day delivery plan of international aviation companies, such as Boeing and Airbus, the average delivery cycle is as long as three months, which lags behind the international advanced level. In the long run, this will hurt the overall reputation of the company and will not be conducive to the economic development of the company. In order to improve delivery efficiency, it is urgent to design an incentive contract to encourage various departments to actively handle delivery problems and improve the efficiency of problem handling.
   Based on this, the management philosophy of “high quality, punctuality and low cost” of complex weapon equipment manufacturing enterprises is integrated into the research. Relying on the extended multi-task dual-agent principal-agent model, the principal-agent relationship between a principal (the main management department of complex weapon equipment delivery, referred to as the delivery center in this paper) and two agents (departments that can deal with delivery problems during the delivery process) is considered. The Cobb-Douglas production function is used to represent the output benefits, which is close to the current situation of “single task failure to meet the standard leads to delayed delivery” in practice, and the relative importance and complete irreplaceability of multiple tasks are balanced, which makes up for the defects of linear output expression. At the same time, comprehensively considering the emotional factor, the risk factor, and complex relationships, the multi-task delivery incentive model has been constructed, which takes the delivery problem handling cycle, quality, and cost as tasks. The study will explore the design method of incentive contracts when the fairness preference agent relationship is neutral, competitive, and cooperative. And the influence of various factors on the effectiveness of incentive contracts is analyzed so that the fairness preference agent can be motivated to actively deal with problems in different relationships to maximize the delivery utility.
   The study shows that the competitive and cooperative relationship between technical business departments promotes the exposure of effort-input information, which can reduce the degree of information asymmetry. The delivery center can promote departments to deepen collaboration to improve delivery efficiency. The multidimensionality of delivery tasks requires the delivery center to consider the relative importance in a balanced way when making decisions. The strategic goals of enterprises change with the development of national strategies and social and economic environment, and the focuses at different stages are also different. When making decisions, each department tends to prefer tasks with relatively high importance and low effort costs so the delivery center needs to balance the cost and importance of each task and drive each department to invest more in important tasks. Incentive contracts can be designed according to the development goals of enterprises at different stages. Conservative technical business departments will take a wait-and-see attitude and lack innovation in organizational changes and business transfers, so they will be distinguished from adventurous technical business departments when incentivizing them. When the relationship between departments is different, the incentive contract will be different. The appropriate incentive contract can be selected according to the type of problem. The optimal incentive coefficient is affected by the department’s factors and there is a weak incentive zone. The internal perception and external environment of the delivery link will lead to changes in the relationship between departments. Therefore, the delivery center can appropriately adjust the delivery environment to achieve the incentive effect. Finally, the case study of the paint quality problem of the aircraft support equipment cockpit ladder in the delivery process of an aviation weapon equipment development and production company is used to verify the conclusions of this study so as to provide theoretical support for the design of incentive contracts in the delivery process of complex weapon equipment. However, the study still has certain limitations. It can be expanded in the direction of the correlation of multi-task effort costs and the dynamic change in task importance in the future.

Research on the Impact of Digitalization on Environmental Service Enterprises’ Performance ——Am Empirical Research Based on Annual Report Text

HU Dongbin, ZHOU Pu

2024, 33(8):  15-22.  DOI: 10.12005/orms.2024.0245

Asbtract ( )   PDF (1029KB) ( )  

References | Related Articles | Metrics

Under the background of “double carbon”, it is of great importance to balance economic development and environmental protection. The development level of environmental service enterprises, as market players of environmental governance and ecological protection, has a direct impact on regional environmental quality. It is very important to reveal the development law of environmental service enterprises for China to promote green development strategy and realize the goal of “double carbon”. Digital transformation is the key to improving service quality and competitiveness of environmental service enterprises. But in the process of practice, enterprises generally face the problems with “not wanting transformation”, “fearing transformation” and “being unable to transform”. Therefore, it is of great practical significance to evaluate the economic effect of digital transformation scientifically and explore the mechanism of enterprise digital transformation. A review of the existing literature reveals that research on digitalization has focused on the macro level such as society and regions. Besides, there are relatively few studies from the microscopic perspective at the enterprise level. Only a few studies show that digital transformation can improve the innovation performance, production efficiency, and organizational authorization behavior of enterprises. Improving enterprise performance is the fundamental purpose of enterprise operation. However, the relationship between digitalization level and enterprise performance is still controversial and lacks the support of empirical evidence, and its internal influence mechanism is not clear. Based on this, this paper constructs a theoretical framework from three dimensions: direct impact mechanism, indirect impact mechanism and heterogeneous impact mechanism of digital transformation on enterprise performance, and attempts to clarify the law of the impact of digital transformation on enterprise performance. The possible contributions of this paper are as follows: First, limited by the fact that it is difficult to measure digitalization at the micro level, the existing literature mostly stays on theoretical analysis and case studies. In this paper, the text analysis technology is used to construct digital transformation indicators, and the panel data model is constructed to fill the existing gap in the empirical study of environmental service enterprises’ digital transformation. Second, based on the realistic background of the servitization transformation of environmental service enterprises, the patterns of the impact of digital transformation on enterprise performance under the scenario of environmental service model innovation are explored, which effectively enriches the situational research perspective of the relevant literature. Third, this paper empirically tests the economic effects of the digital transformation of environmental service enterprises, constructs a relatively systematic theoretical framework of “digital transformation—enterprise performance”, which is conducive to opening the mechanism “black box” between digital transformation and enterprise performance of environmental service enterprises, and providestheoretical guidance for the digital transformation of environmental service enterprises.
   This paper takes Chinese listed environmental service enterprises from 2010 to 2019 as research samples, uses the text analysis technology to construct digital transformation indicators, empirically tests the economic effects of digital transformation on environmental service enterprises, explores the indirect transmission mechanism of digital transformation on enterprise performance, and further analyzes the heterogeneous effects under different enterprise characteristic scenarios. The study has found that: (1)From the direct effect, digital transformation has a significant positive impact on the economic performance of environmental service enterprises, and the driving effect lasts for a long time. (2)The indirect transmission mechanism test shows that operating efficiency directly affects enterprise performance, while digital transformation can indirectly improve enterprise performance by improving operating efficiency of environmental service enterprises. (3)From the perspective of heterogeneity of service model, the economic effect of digital transformation is significant for enterprises implementing new service model, while for enterprises still using traditional service model, the impact of digital transformation on enterprise performance is not significant. Therefore, the service model has a heterogeneous effect on the economic effect of digital transformation. (4)From the perspective of enterprise ownership, digital transformation has a significant promoting effect on the performance of non-state-owned enterprises, while the impact of digital transformation on the performance of state-owned enterprises is not significant. Therefore, the enterprise ownership has a heterogeneous effect on the economic effect of digital transformation. This paper contributes to clarifying the impact effect and mechanism of digital transformation on enterprise performance, and also provides useful insights to drive digital transformation in environmental service enterprises.

Rural E-commerce Logistics Route Optimization by Joint Truck and Multi-drone Delivery Considering Waiting Cost

CHEN Xiqiong, WANG Xinglong, HU Dawei

2024, 33(8):  23-29.  DOI: 10.12005/orms.2024.0246

Asbtract ( )   PDF (1307KB) ( )  

References | Related Articles | Metrics

The emergence of logistics drones provides a new way to solve the problems with rural e-commerce logistics, such as scattered demand points, road network limitation, and high terminal delivery costs. However, the terminal delivery by drones is limited by the cruising range and load capacity, which restricts its wide-scale application. In order to expand the service scope of logistics drones and reduce the cost of terminal distribution, the joint distribution mode of trucks and drones is a brand new attempt. This paper studies a Travel Salesman Problem with multi-Drones (TSP-mD), in which a truck with several drones is required to visit a set of customers. The truck starts from a depot and back to the depot after all customers are visited by the truck or a drone once. This study is of great significance for reducing the cost and improving the efficiency while applying to the rural logistics scenario.
   Considering the mutual waiting cost of trucks and drones, this paper establishes a joint delivery route optimization model for single-truck and multi-drone with the goal of minimum total cost. According to the characteristics of the model, an adaptive large-neighborhood search algorithm is designed. Three damage operators and repair operators are used, and the simulated annealing acceptance criterion is used for the temporary solutions after damage and repair. Based on the proposed algorithm, two types of instances (9 under each type) including 10-100 nodes with uniform distribution and cluster distribution are solved. The comparison with the CPLEX solution results indicates that the proposed algorithm has faster speed, better accuracy and stability. Through the analysis of the joint delivery scheme of trucks carrying different numbers of drones, the results show that the delivery scheme of truck carrying 0 drones (pure truck delivery) and the joint delivery of truck carrying 1-4 drones can reduce the total rural e-commerce terminal delivery cost by 17.88%~28.89%, and can improve the delivery efficiency. Finally, the sensitivity analysis of the cruising range of the drones shows that with an increase in the cruising range, the delivery cost decreases first rapidly and then slowly or tends to be stable.
   For future research, the time window constraint would be considered, and the influence of factors such as load constraint, geographical conditions, weather impact, fixed cost of carrier and personnel cost would be taken into account. Besides, the joint distribution mode of multiple trucks and multiple drones is a promising research direction to further improve the distribution efficiency.

Research on Dynamic Productivity Planning and AGV Configuration of Photovoltaic Cell Production Workshop

LI Kunpeng, HAN Xuefang

2024, 33(8):  30-36.  DOI: 10.12005/orms.2024.0247

Asbtract ( )   PDF (1237KB) ( )  

References | Related Articles | Metrics

For the past years, as the market size of the photovoltaic industry has continued to expand, the demand for battery cells as the core component of photovoltaic power generation has increased significantly. However, the production technology of photovoltaic cells is complex, involving multiple processes and precision machinery. Photovoltaic cell manufacturers are generally challenged by high costs and low efficiency. On the one hand, in the photovoltaic cell manufacturing workshop, different processes have different numbers of machines and processing speeds, and the maximum productivity of the workshop is limited by the productivity of the bottleneck process. It is a challenge for the photovoltaic cell production workshop to balance the productivity of machines in different processes and ensure the maximum production efficiency of bottleneck processes. In addition, due to a large number of machines in the workshop, it is difficult to repair and maintain them in a timely manner, and the workshop inevitably encounters unexpected accidents such as machine breakdowns. This may cause the bottleneck process to shift, which in turn causes production reduction, production stoppage and other problems, resulting in reduced workshop productivity and wasted production resources. Therefore, after the machine breakdown, adjusting the production speed of other machines is particularly important. But in fact, most enterprises still rely on manual to plan the workshop productivity. This method lacks science and timeliness, and cannot meet the enterprise cost reduction and efficiency of the real needs. On the other hand, with the promotion and implementation of intelligent manufacturing, automatic guided vehicle (AGV) as an important symbol of factory intelligence, has become the main material handling tool in the production workshop of photovoltaic cells. AGVs have the advantages of high efficiency and accuracy, but their purchase cost is also a major challenge for most companies. Therefore, making a reasonable decision on the number of AGVs to be purchased has become another challenge for photovoltaic cell manufacturing enterprises.
   This paper studies a dynamic productivity planning and AGV configuration problem in a photovoltaic cell manufacturing shop based on the actual production problem of a typical photovoltaic cell manufacturing company. The problem considers the characteristics of fixed capacity caching, random machine failures, and AGV handling materials. In order to solve the problem, firstly, this paper establishes a mixed integer linear programming model with the objective of maximizing the daily productivity of the workshop. Secondly, we propose two cache usage strategies based on the constraint theory, namely, “prioritizing the use of cache and prioritizing the consistency of the productivity of each process”, and design the corresponding productivity dynamic planning algorithms respectively. Meanwhile, this paper proposes a control algorithm to ensure that the productivity of all processes is always consistent with the productivity of the bottleneck process, which is a commonly used productivity planning scheme in current production workshops. Furthermore, since AGVs are a key component of production and cost, this paper introduces an AGV allocation algorithm. This algorithm is employed to calculate the optimal number of AGVs under different production conditions. Finally, in order to compare the performance of the two productivity planning schemes proposed in this paper with the actual schemes adopted by enterprises, we extract the actual workshop data, and set up three schemes, namely, “Prioritizing the use of caching”, “Prioritizing the maintenance of productivity consistency”, and the no-caching strategy, with different breakdown scenarios. The experiments are conducted under different breakdown scenarios to explore the changes in productivity and machine working time. The experimental results show that the strategy of having cache zones and giving priority to the use of caches can help enterprises better cope with unexpected situations such as machine breakdowns. In addition, through the simulation experiments, this paper calculates the optimal number of AGV configurations for the three scenarios when there is no machine breakdown. Moreover, the effect of AGV handling time on the productivity and the number of AGV configurations are further discussed. The research results can not only help the cell manufacturing workshop to plan the production productivity in real time and decide the AGV configuration scheme, but also can be extended to other workshop production and logistics planning scenarios with productivity imbalance.

Research on Online Production Scheduling of Steel Box Girder Section with Degradation Effect

YANG Xiaohua, MA Ran, ZHANG Yuzhong

2024, 33(8):  37-43.  DOI: 10.12005/orms.2024.0248

Asbtract ( )   PDF (1173KB) ( )  

References | Related Articles | Metrics

As a basic component of bridge assembly, a steel box girder section is in great demand, and the optimization of its production plan has a great influence on manufacturers, so the research on online production scheduling of the steel box girder section becomes more and more urgent and important. It has been shown that in the production scheduling of steel box girder sections, there is relatively little research on the online scheduling problem considering the degradation effect. Based on the existing research, this paper considers the single machine scheduling problem of steel box girder segment online production with degradation effect, aiming at minimizing the maximum weighted completion time, presents an online algorithm for the studied model, and implements the simulation. On the one hand, the model of online scheduling is perfected, which fills the blank area of existing research. On the other hand, the different weights of steel box girder sections and the degradation effect of processing are considered simultaneously, which provides a strategic reference for more realistic production scheduling problems and brings reliable selection of optimization schemes for managers.
   The two online scheduling problems studied in this paper consider the job weight and processing degradation effect, and minimize the maximum weighted completion time as the optimization goal. It is worth mentioning that we realize that the waiting strategy can reduce the losses caused by delaying orders with more weight that will soon be available, so we propose a deterministic online algorithm in polynomial time based on the waiting strategy, which is suitable for both problems studied. For the first problem studied, the processing time model is p_j=α_jt(t>0). First of all, it is concluded that the problem with the lower bound is 1+α_max. Then we prove offline optimal sorting methods for the problem, and design online algorithm H₁. Lastly we prove H₁ is the best possible online algorithm with the competitive ratio of 1+α_max by stepwise analysis. In response to the second question studied, the processing time model is p_j=α_j(A+Bt)(A>0). In using the same technique to draw the lower bound of the problem of 2+Bα_max, we prove offline optimal sorting methods for the problem, then design and analyse online algorithm H₂, and H₂ is the best possible online algorithm with the competitive ratio of 2+Bα_max.
   In conclusion, this paper gives the best possible online algorithms for both models studied. In view of the theoretical results obtained, this paper further carries out numerical simulation, using Python software 3.9 version to implement it. In numerical simulation, for the first problem Γ₁, by the application of randomly generating jobs instance of size n∈ {5,10,30,50,100,150,200,300,400}, respectively for 500 times and 1000 times of experiments, it is concluded that the average and maximum of performance ratio under each array are not more than the competitive ratio theoretically provided, so the correctness and effectiveness of online algorithm proposed in the paper are proved. For the second problem Γ₂, considering the changes in parameters A and B, we randomly set several groups of A and B where the number of regenerated jobs is n∈{10,50,100,200,300,400}, and conduct 500 and 1000 experiments respectively. It is concluded that the average value of the ratio of performance ratio to competition ratio under any array is not more than 1. This also illustrates that the correctness and effectiveness of online algorithm proposed in the paper are proved.
   In the future, it is worth expanding our research from several aspects for reference in the future work. In the actual production of steel box girder sections, a group of jobs of the same type may be placed in the same batch for processing, so the online scheduling of batch processing is worthy of attention. At the same time, in the face of large quantities of steel box girder segment processing, processing equipment may also be composed of more than one machine, so the parallel machining of the steel box girder segment is also the focus of the next research direction. Further, we can study the joint scheduling of product processing and distribution, which is also a very influential research direction.


Dynamic Process Quality Anomaly Recognition Method Based on Feature Reduction and Improved Support Vector Machine

LIU Li, LIU Yumin, ZHAO Zheyun

2024, 33(8):  44-50.  DOI: 10.12005/orms.2024.0249

Asbtract ( )   PDF (1456KB) ( )  

References | Related Articles | Metrics

Industrial equipment operating at high speeds generates a large amount of dynamic data streams that reflect fluctuations in its production process and operational status. However, due to the complexity of dynamic process big data, it is difficult to achieve a high recognition accuracy for any type of abnormal pattern relying solely on a single type of data feature. The urgent problem to be solved is how to fully integrate, screen, and efficiently utilize information. At present, intelligent technology has become one of the effective methods to solve complex production processes with characteristics such as nonlinearity, multiple inputs, and multiple outputs, due to its advantage of not requiring the establishment of precise mathematical models. Therefore, classifiers used for dynamic process quality anomaly pattern recognition have evolved from early rule-based expert system judgment to the combination of support vector machines, particle swarm optimization algorithms, rough sets, and other technologies. The RS-GA-SVM model proposed in this article is a computationally simple and highly accurate method for identifying quality anomaly patterns, which is beneficial for real-time quality monitoring and fault diagnosis in automated manufacturing industries such as petroleum and chemical.
   In order to effectively reduce the dimensionality of features and improve the recognition accuracy of abnormal patterns in dynamic processes, this paper combines feature reduction with improved support vector machines to identify abnormal patterns in dynamic process quality. Firstly, it uses Monte Carlo simulation to generate a dynamic data stream of the production process, including training samples and test samples. The test samples are mainly used for estimating the performance of SVM. Next, it extracts 16 statistical features and 7 shape features that can characterize quality anomaly patterns from the raw data. Secondly, this article uses rough sets (RS) to reduce the feature set, eliminate redundant and interfering features, and obtain the optimal attribute set. Subsequently, it uses genetic algorithm (GA) to find the optimal parameters for support vector machine (SVM), which can reduce the subjectivity of SVM in parameter selection. Finally, it uses the GA-SVM model to identify quality anomaly patterns, and compares the recognition accuracy with other similar models to verify the effectiveness and applicability of the proposed recognition model.
   The results show that the statistical features and shape characteristics used in this article can combine multiple quality mode features, integrating the classification advantages of different features on different quality modes. Secondly, the 12 features obtained after rough set filtering have strong ability to distinguish whether there are abnormal conditions in dynamic processes, which can effectively reduce the input dimension of support vector machines. Finally, this article uses simulation experiments to compare the recognition accuracy with other models of the same type. The results show that the genetic algorithm optimized support vector machine has significantly higher accuracy in identifying quality anomaly patterns than other models, indicating that the RS-GA-SVM model proposed has good recognition accuracy and robustness, and can effectively monitor dynamic processes.
   In future, due to uncertain factors in the actual production process that may lead to the emergence of new quality anomaly patterns or multiple concurrent quality patterns, it is necessary to conduct in-depth research on these unknown quality anomaly models. Meanwhile, this article only uses the equal frequency interval partitioning method to reduce the combination of serial features, and rough sets also include other attribute reduction methods. Therefore, there is still a lot of research space for feature reduction methods based on rough sets.

Hybrid Estimation of Distribution Algorithm for Distributed Flexible Job Shop Scheduling

WEI Guangyan, YE Chunming

2024, 33(8):  51-57.  DOI: 10.12005/orms.2024.0250

Asbtract ( )   PDF (1421KB) ( )  

References | Related Articles | Metrics

In the context of economic globalization, many manufacturing enterprises are being guided by new manufacturing modes to establish distributed manufacturing units, aimed at cost savings and enhancing regional competitiveness. The scheduling decisions of manufacturing systems are transitioning from a centralized single-node model to a distributed multi-center approach. At the same time, small-batch, diverse, and personalized manufacturing services are promoting the development of flexible manufacturing units. The scheduling demands for distributed flexible manufacturing systems have garnered widespread attention. Therefore, this paper addresses the Distributed Flexible Job Shop Scheduling Problem (DFJSP), proposes a model that optimizes total cost and tardiness, and designs an H-EDA-TS algorithm combining estimation of distribution algorithm and Tabu search for solving the model.
   Considerations of cost and time disparities among different machines in heterogeneous factories are incorporated into this study, which formulates a DFJSP model with the objective of minimizing total cost and total tardiness. Several constraints are integrated into the model to simulate practical manufacturing conditions: (1)Each job is allocated to a single factory.(2)Multiple jobs can be assigned to each factory. (3)Operations are uninterrupted, where completion times are equal to start times plus durations. (4) The completion time of a job’s final operation defines its departure time from the factory. (5)Machines can process only one job at any given time. (6)Each operation is exclusively performed by one machine at a time. (7)Operations adhere strictly to their designated routing sequences. (8)All operations for each job must be fully executed.
   Distributed flexible job shop scheduling problem is comprised of three sub-problems: allocating jobs to factories, sequencing job processing within each factory, and selecting suitable machines for operations. This paper introduces a three-dimensional encoding scheme for the DFJSP model to represent the solutions. Each layer of the encoding corresponds to a sub-scheduling problem. Then, an H-EDA-TS algorithm is devised, which combines the global search advantages of estimation of distribution algorithm with the local search advantages of the Tabu search. The algorithm consists of an EDA component with three probabilistic models for population sampling and a TS component with five neighborhood structures designed to optimize objectives. Additionally, to conserve computational resources, an adaptive mechanism based on the sigmoid function regulates the initiation conditions and frequency of the TS component. Since this paper considers two optimization objectives, namely, minimizing total cost and total tardiness, the Pareto optimality principle and the crowding distance operator from NSGA-II are employed for comparing and selecting solutions. Finally, experiments conducted on various scales of DFJSP instances demonstrate the significant advantages of the proposed H-EDA-TS algorithm over the estimation of distribution algorithm, Tabu search, and hybrid estimation of distribution algorithm and variable neighborhood search algorithms.
   With growing environmental awareness and governmental support for green manufacturing policies, manufacturing enterprises are increasingly focusing on energy savings and emissions reduction during production processes. Carbon emissions can serve as an optimization target in scheduling decisions, aiming to achieve green production through the selection of low-energy-consumption scheduling schemes. Additionally, ensuring equitable workload distribution among manufacturing units is a noteworthy scheduling objective. Further investigation is warranted to refine scheduling optimization algorithms that holistically consider economics, time, environment and fairness. Furthermore, this paper addresses the static scheduling problem of distributed flexible job shops. However, as smart factories become prevalent, dynamic DFJSP research presents an intriguing avenue for future exploration.


Integrated Optimization of Makespan, Robustness and Energy Cost for the Flow Shop in Manufacturing Plant

CUI Weiwei, JIANG Chengren, LIU Xinbo

2024, 33(8):  58-64.  DOI: 10.12005/orms.2024.0251

Asbtract ( )   PDF (1000KB) ( )  

References | Related Articles | Metrics

Managers within manufacturing plants confront increasingly intricate scenarios, necessitating efforts to minimize manufacturing lead times amidst the destabilizing impact of random failures. Concurrently, they must also endeavor to curtail energy costs within time-of-use tariffs, thereby bolstering the price competitiveness of products. Focusing on the discrete flow shop, this study incorporates considerations of energy consumption costs within the framework of TOU tariff policies and the stochastic nature of equipment failures. Through the integration of production scheduling and equipment maintenance, this study aims to devise a cohesive modeling approach that enables comprehensive planning for both activities. The devised integrated optimization scheme outlined in this paper is poised to significantly aid enterprises in achieving peak shaving and valley filling, alongside cost reduction and efficiency enhancements under time-sharing tariff policies. Furthermore, it offers valuable insights for workshop managers seeking to formulate judicious and effective production plans within complex and uncertain production environments.
   This study focuses on multiple interrelated dimensions within the manufacturing shop, establishing a mathematical model that encompasses three decision dimensions: production scheduling, equipment maintenance, and energy allocation. A two-layer algorithm is devised to tackle this model effectively. Firstly, a method based on surrogate measures is designed to evaluate the performance of solutions. Then, a metaheuristic algorithm is designed combining the NSGA-II framework and the constructive-heuristic rules to search the Pareto curve of this multi-objective problem. Data are acquired through Monte Carlo simulation. Under the assumption that random faults follow an exponential distribution, random numbers are generated by sampling iteratively to simulate the system, followed by conducting several tests.
   In the algorithm’s validation phase, Monte Carlo sampling simulation is employed to compute the expected value of the objective function within the inner algorithm. Subsequently, an appropriate proxy index function is devised to effectively approximate the expected target value, significantly reducing operational time while maintaining a certain level of precision. Comparative analysis of VEGA reveals substantial enhancements in the robustness index of the NSGA-II algorithm, formulated within the outer algorithm, along with improvements in product delivery time and electricity cost. The intricate strategy proposed herein for model validation effectively mitigates electricity costs. In essence, the incorporation of buffer time insertion and the algorithm outlined in this study enhances system performance concerning stability and electricity cost indices when encountering random faults. Our findings demonstrate the efficacy of buffer time insertion: it mitigates fault impacts on subsequent processes, ensuring current process stability, and diminishes the proportion of processing time subject to peak electricity prices, thus economizing on electricity costs. Additionally, our investigation indicates that an increase in buffering time leads to a gradual reduction in peak power, further curtailing total electricity costs. However, practical implementation may be constrained by extended delivery times, rendering this approach less advisable in practice.
   In future research, the problem can be extended by addressing the following two aspects: (1)Assessing the influence of renewable energy power supply modes on all facets of the system. (2)Investigating pertinent issues within alternative production layouts, such as flexible assembly line shops, open shops, and others.

Electric Vehicle Charging Station Location Optimization Based on the Valet Charging Mode

TONG Min, HU Zhihua

2024, 33(8):  65-71.  DOI: 10.12005/orms.2024.0252

Asbtract ( )   PDF (1671KB) ( )  

References | Related Articles | Metrics

According to the statistical analysis of the China Association of Automobile Manufacturers (CAAM), in March 2022, the production and sales of new energy vehicles in China reached 465,000 and 484,000 units, respectively, with an increase by 25.4% and 43.9% in the chain and an increase by 1.1 times in both year-on-year, and the market share reached 21.7% and continued to maintain the momentum of rapid growth. However, the problems with hard-charging, long charging time, and frequent queues often make users of new energy vehicles face the dilemma of being unwilling to charge and having no time to do so. To solve these problems, a new service model has emerged, namely the valet charging service model. Users can make an appointment through the Valet Charging App, and the charging staff will arrive at the user’s location within a specified time, drive the user’s vehicle to a nearby charging station, complete the charge and return to deliver the vehicle to the user. The research on the valet charging service model is of great importance to alleviate the pressure of urban charging, reduce the charging and waiting anxiety of users, and explore the potential users of new energy vehicles.
   In recent years, the problem with optimizing the distribution of the electric vehicle (EV) charging stations has become a hot topic of research for domestic and foreign scholars. However, there are still few studies on the location problem with charging stations under the valet charging service mode. LI et al.(2021) considered the stochastic nature of user charging demand and modeled the problem using the Two-stage Stochastic Mixed-Integer Program (TSMIP). Based on their work, the rate of users choosing valet charging service is calculated by integrating the charging time cost and the valet charging service price, and the TSMIP model is developed by combining this parameter to minimize the construction cost of the charging infrastructure and the travel cost of EV users to the charging station while maximizing the profit of the valet charging business. The Sample Average Approximation (SAA) and Epsilon constraint method are used to solve the multi-objective TSMIP model to explore the impact of the location strategy charging station on the operation of the valet charging service.
   Based on the results of numerical experiments, we find that while maximizing the expected profit of the valet charging business with a fixed number of charging stations, the location of charging stations will gradually spread around the region, increasing the cost of charging travel for EV users. Therefore, decision-makers should take into account the actual situation and the interests of both enterprises and users in choosing the location strategy of charging facilities. Moreover, the upper bound of charging cost per unit of time has a more significant impact on the expected profit of the valet charging business than the lower bound of cost per unit time of charging for users. Therefore, when conducting market research in a region, companies should thoroughly investigate and obtain information about the upper bound of cost per unit of time. This parameter can be initially estimated using the service price ratio to the region’s minimum charging time and then adjusted according to the actual situation.
   Valet charging is mainly included in the after-sales services of large new energy vehicle companies. However, given the rapid growth of new energy vehicle users and charging demand in recent years, the development of valet charging mode as a stand-alone business has broad prospects. This paper proposes the following directions for future research on issues related to the valet charging service model: the problem with scheduling optimization and route optimization for charging staff and the problem with optimal pricing for valet charging service.

Robust Optimization of Location-Resource Allocation Integrated Decision for Blood Collection Points under Uncertain Environments

ZHOU Yufeng, HU Huanqing, CHEN Liangyong

2024, 33(8):  72-78.  DOI: 10.12005/orms.2024.0253

Asbtract ( )   PDF (974KB) ( )  

References | Related Articles | Metrics

In recent years, the blood shortage in mainland China has become increasingly serious, endangering people’s lives and health safety. In order to increase the blood collection volume and alleviate the blood shortage, it is necessary to improve the layout system of blood collection facilities, optimize the allocation of blood collection resources, and construct an efficient and reliable blood collection network. There are two types of blood collection facilities, fixed blood collection houses and mobile blood donation vehicles. Blood collection houses are permanent facilities that, once established, are difficult to change in the short term. Blood donation vehicles are mobile facilities whose location decisions are variable from one period to the next. Therefore, the problem is defined as a dynamic optimization problem of location-resource allocation integration for two types of heterogeneous blood collection points. Therefore, this study proposes a heterogeneous blood collection point location-resource allocation integrated decision optimization problem under uncertain environments. The problem addresses the following decisions. Where should fixed blood collection houses be located? How to dynamically locate blood collection points for mobile blood donation vehicles? How to dynamically allocate resources such as blood collection personnel and equipment to blood collection facilities?
   China’s blood collection practice shows that the number of volunteers and blood supply at blood collection points cannot be obtained directly. Comprehensive assessment of candidate locations for blood collection vehicles and blood collection houses by means of key indicators and then determining the coverage weights will make the decision of optimizing the location of blood collection points more reasonable and operational. These key indicators include: the flow of people in the target area, the accessibility of people, the level of blood collection service, the activity of blood donation, and the interval between blood collection periods. A comprehensive multi-attribute evaluation of the above key indicators can obtain their coverage values. In this study, the generalized maximum coverage model is introduced to establish an integrated decision-making model of blood collection point location-resource allocation with the goal of maximizing the coverage weight. Subject to the constraints of realistic factors, the number of available blood donation vehicles, personnel and equipment available for allocation in each period are uncertain. Robust optimization techniques are used to deal with the uncertainties in the parameters of available blood donation vehicles, personnel and equipment. The corresponding robust optimization model is developed based on linear programming dyadic theory.
   Aiming at the characteristics of the model, an improved grey wolf optimization algorithm (IGWO) is designed to solve the model. The improvement strategies are as follows. Firstly, the dynamic weighting factor is introduced to accelerate the convergence speed of the algorithm and improve the optimization performance. Secondly, the simulated annealing Metropolis criterion is introduced to prevent the algorithm from falling into premature maturity. Thirdly, the 3-opt local optimization strategy is introduced. The above improvements better balance the exploration and exploitation capabilities of the algorithm.
   Different cases with different sizes are set up to test the performance of IGWO and compare it with the traditional grey wolf optimization algorithm (GWO) and particle swarm optimization algorithm (PSO). The results show that IGWO has obvious advantages in convergence speed, solution accuracy, and stability of solution compared with traditional GWO and PSO, which proves the effectiveness of the proposed algorithm.
   The results also show that although the robust optimization makes the total coverage of the localization points decrease, the decrease ratio is small, in which the optimal value gap is only 0.19%. Thus, robust optimization reduces the risk caused by uncertainty. In a realistic decision-making environment, the uncertainty of the parameters should be fully considered for robust optimization of the positioning and resource allocation decisions of blood collection points to reduce the risk of uncertainty.

Robust Optimization of Logistics Network Considering Low-carbon Policy and Uncertain Demand

JIANG Jiehui, SHENG Dian

2024, 33(8):  79-85.  DOI: 10.12005/orms.2024.0254

Asbtract ( )   PDF (1537KB) ( )  

References | Related Articles | Metrics

In response to global warming, the Chinese government will adopt more effective policies and measures, which aim to peak carbon dioxide emissions by 2030 and achieve carbon neutrality by 2060. Such an ambitious goal has brought great challenges to various industries by changing the mode of operation management and implementing the transformation of the low-carbon economy. In terms of the total carbon emissions in China, the transportation sector currently accounts for about 10.4%, of which the road is the main carbon emission contributor, accounting for 87%. It is an important and necessary way to achieve the emission reduction target through the optimization and reconstruction of the cargo transportation system. In terms of government regulation policies, the mostly discussed are carbon emission cap-and-trade and carbon tax. They are set to internalize the cost of carbon emissions, encourage enterprises to actively optimize their logistics network structure, and promote the transport mode from road to cleaner rail and water transport. China began implementing a carbon cap-and-trade policy in eight pilot cities in 2011, and established a carbon emission trading market in 2021. Therefore, the integration of carbon cap-and-trade policy in the design of supply chain logistics networks and the construction of multi-mode transport systems are effective ways for enterprises to transform low-carbon.
   The classic facility location problem focuses on the goals of cost, efficiency, and social welfare improvement while ignoring the environmental impacts of transportation activities. The design of the supply chain logistics network has a significant two-stage property: during the effective period of facility planning decisions, the future product demand is changed by economic and social development and affects the decision of the enterprise. In recent years, probabilistic distributed fuzzy sets have been widely used to reduce their conservatism and improve their robustness by constructing fuzzy sets according to confidence levels from probability distribution information in historical data. In addition, the robust optimization method is an effective method to solve the optimization problem considering uncertainty. Relevant research provides a comprehensive optimization framework for supply chain logistics network design, but this method does not involve how to reconstruct the distributed robust model into a standard programming model that can be directly used by commercial solvers, which can design an efficient solution.
   Therefore, this paper studies the design of a supply chain logistics network under product demand uncertainty and carbon cap-and-trade policy. Considering the uncertainty of future product demand, a norm-based fuzzy set is proposed to describe the probability distribution of uncertain demand. Given the carbon emission reduction and trading policies, a two-stage robust optimization model for supply logistics network design is constructed to optimize the number and scale of distribution centers and the multi-mode transportation scheme of products. Based on the classic three-level network planning, the model comprehensively considers the distribution center construction cost, transportation cost, product shortage cost, and the negative utility of carbon emission trading to minimize the total cost. Considering the separability of the model, this paper reconstructs the main model and submodels and proposes a solution method based on column and constraint decomposition. Finally, the validity of the proposed robust optimization model and decomposition algorithm is verified by a practical case.
   The numerical results show that the proposed distributed robust optimization method can improve the robustness of the decision scheme and reduce its investment effectively. The greater the demand uncertainty, the greater the number and scale of distribution centers expected to be built, which in turn increases the corresponding expected construction costs, transportation costs, carbon acquisition, and total costs. In order to realize the carbon emission reduction target, government departments can encourage enterprises to implement low-carbon production planning and operation in combination with policies such as subsidies for low-carbon transport modes. It is also important to strengthen the supervision of carbon emissions and assess the emission reduction performances under different incentive measures. When the carbon quota is abundant, enterprises should pay more attention to reducing the transportation cost. And when the quota is scarce, the focus will be on how to reduce transport carbon emissions.

Research on Retail Logistics Collaborative Scheduling Optimization Considering Differentiated Service Time

LI Wenli, TIAN Qiannan, HE Peiyang, WANG Xiaoyan, GUO Hao

2024, 33(8):  86-92.  DOI: 10.12005/orms.2024.0255

Asbtract ( )   PDF (1022KB) ( )  

References | Related Articles | Metrics

The epidemic has accelerated the expansion of traditional retailers to online businesses and the emergence of new retail models integrating online and offline, making the timeliness of retail logistics and service quality the key to winning the market in the post-epidemic era. Under the new retail model, the logistics demands of consumers are more dispersed and the batch frequency is higher. Different orders and logistics distribution need to be uniformly deployed, which puts forward higher requirements for the service quality and timeliness of logistics distribution. Retail logistics distribution service requires the cooperation among logistics personnel, distribution vehicles and customers, ignoring the differentiated service time of logistics personnel, the different departing time of vehicles due to different release dates, and the timeliness of customer demand lead to high cost, poor timeliness and low customer satisfaction of retail logistics distribution schemes. When logistics personnel are familiar with the surrounding environment of customers, they can quickly and accurately find the distribution address and appropriate parking points, and even get familiar with the traffic conditions to avoid the congested sections. Especially in the post-epidemic era, the epidemic situation presents a multi-point distribution state, customer demands are scattered and change every day, and logistics distribution personnel are also absent due to temporary lockdown. If differentiated service time is ignored, this will not be conducive to the full utilization of human resources, or to the response to emergencies. Order release dates determines the vehicle departure and distribution time, which not only affects the collaborative decision of order assignment and route planning, but also has an important impact on distribution efficiency. In addition, as the distribution process is often accompanied by force majeure factors such as bad weather and traffic jams, the customer delivery time window is usually flexible, which means that the customer is allowed to receive the service earlier or later than the specified time window to a certain extent, but the violation of the specified time window will lead to the decline of customer satisfaction, so the corresponding punishment is considered in the objective function.
   In this context, we study the collaborative scheduling problem of retail logistics aiming to minimize travel cost, penalty cost and differentiated service time cost. Under the constraints of order release dates and customer flexible time window, dispatching a group of logistics distribution personnel to complete the delivery tasks of customer orders is a collaborative optimization of task assignment and vehicle routing planning for multiple logistics distribution personnel with differentiated service time.To solve this problem, a linear mathematicalprogramming model is established with the optimization objective,and the improved iterated local search algorithm based on large neighborhood search process is designed in this paper. This algorithm uses regret repair operator to generate high quality initial solution to enhance the search efficiency, and introduces large neighborhood search with four removal operators and two repair operators, a breaking mechanism and optimal service start time model to enhance the global optimal search ability of the algorithm.
   Finally, the numerical experiment verifies the effectiveness of the model and algorithm by solving benchmarking instances and the numerical instances in the paper, and the sensitivity analysis of the parameters gives corresponding management enlightenment, which provides effective reference and suggestions for the effective management of distribution personnel, efficiency improvement of distribution and cost control of retail logistics operation management in the post-epidemic era.

Dynamic Risk Management Optimization of Online Car-hailing in the Sharing Economy

LI Na, MA Deqing, HU Jinsong

2024, 33(8):  93-100.  DOI: 10.12005/orms.2024.0256

Asbtract ( )   PDF (1186KB) ( )  

References | Related Articles | Metrics

As an emerging business model, the sharing economy has garnered widespread attention from all sectors of society. It emphasizes the sharing of underutilized assets in a manner that enhances efficiency and sustainability, thereby disrupting innovations across various industries, including the taxi sector. However, due to the heterogeneity of participants’ social strata, the complexity of transaction motivations, and the unpredictability of transaction behaviors, consumers perceive risks inherent in the transaction process, which hinders the value creation of the sharing economy. Consequently, a pressing research question arises regarding how governments and enterprises can effectively collaborate to address risks in the sharing economy to enhance the quality and quantity of transactions, thereby ensuring the robust development of the industry.
   Taking online car-hailing as an example, this paper focuses on two types of perceived risks: safety risk and service risk. Considering the occurrence of emergencies in online car-hailing and the effect of reference service quality among consumers, a dynamic game model of “government+platform” risk management is constructed to precisely describe the decision-making process between the government and platform enterprises during the operation of online car-hailing. This paper represents the safety risk in online car-hailing through the occurrence of random crisis events. Service risk refers to the disparity between consumers’ expected service and actual experience, which leads to the possibility of consumers perceiving poor service quality and regret. The effect of reference service quality is used as a metric to measure service risk. From a dynamic perspective, this paper employs differential game theory to derive the optimal venture investment strategies and performance of both parties under scenarios of government-enterprise cooperation, non-cooperation, and government cost-sharing contracts. It quantitatively analyzes the optimization of risk management in the sharing economy.
   The research findings of this paper indicate the following: (1)Collaborative risk management between the government and platform enterprises can maximize system performance. Government cost-sharing contracts can not only effectively reduce safety risks in online car-hailing but also achieve Pareto improvements in the benefits of both parties. However, increases in risk rates, depreciation rates, and the effect of reference service quality can, to a certain extent, weaken the improvement effect of government cost-sharing contracts. (2)Considering the potential risks inherent in the actual operation of online car-hailing, it is beneficial for the government and platform enterprises to make the most economically efficient venture investment decisions by making planned adjustments based on specific crisis scenarios. Increasing investment in risk management for online car-hailing, when risk rates and depreciation rates are low, can have a positive feedback effect on venture investment construction, fostering a virtuous cycle. In contrast, when predicted risk rates or depreciation rates are high, managers should prioritize reducing investment to offset potential future profit losses. Additionally, compared to rebuilding the reputation of online car-hailing platforms after a crisis, managers should prioritize risk prevention before a crisis occurs. (3)The effect of reference service quality can dampen the enthusiasm of governments and enterprises for investing in safety-related ventures but incentivize enterprises to improve the service quality of online car-hailing. This, to a certain extent, can mitigate the negative impact of crises on system performance. This research enriches and enhances our understanding of risks and their management optimization in online car-hailing within the sharing economy.
   The conclusions drawn provide meaningful insights for promoting the sustainable development of online car-hailing and can be extended to other similar sharing economy models, such as homestays in tourism. Furthermore, the findings offer relevant ideas for governments to design and implement new policies and help establish sustainable transportation systems.

Chance-constrained Robust Large-scale Group Consensus for Probabilistic Linguistic Empathy Network

HAN Yefan, JI Ying, QU Shaojian

2024, 33(8):  101-108.  DOI: 10.12005/orms.2024.0257

Asbtract ( )   PDF (1170KB) ( )  

References | Related Articles | Metrics

Empathy relationship, as a social relationship, objectively exists in some realistic group decision-making problems, so group size becomes an important factor affecting group consensus. It increases the uncertainty of group decision-making problems and leads to a costly and time-consuming consensus reaching process. Therefore, this paper aims to investigate a large-scale group consensus decision-making method in a probabilistic linguistic empathy network environment. A robust cost consensus model is then developed using the chance-constrained robust optimization method during the feedback adjustment process to account for the effects caused by the uncertainty of the unit adjustment cost.
   Considering that it is difficult for people to clearly evaluate their empathy relationship with others in real life, it is more expressed in natural language. We first define a probabilistic linguistic empathy function to evaluate the empathy relationship among decision makers, and thus build a probabilistic linguistic empathy network. Based on such a network, the decision makers’ preferences can be decomposed into their intrinsic preference, representing their true opinion, and their empathetic preference for other decision makers in the network.
   Secondly, a fuzzy C-means clustering method based on preference relationship is used to divide decision makers with high similarity into several subgroups. Since the empathy relationship among decision makers should be used as a reliable indicator to assign weights to each cluster, this paper considers the size, cohesion, and overall empathy degree of each cluster to determine the importance of clusters. In addition, since decision makers in the same cluster have highly similar preferences, the degree of attitudinal empathy based on the empathy relationship is utilized to determine the weight of decision makers within a class.
   When the group consensus degree cannot reach a predefined consensus threshold, to improve the consensus quality, opinion adjustment becomes a natural phenomenon. At this point, an efficient feedback adjustment process needs to be implemented. This paper designs an optimization-based feedback mechanism incorporating a minimum cost consensus model, which is guided by empathetic relationship. Affected by factors such as social experience and educational background, the unit adjustment cost of decision makers from different organizations may present uncertainty. Robust optimization, as a powerful tool for dealing with uncertainty, is often combined into minimum cost consensus models to deal with uncertain unit adjustment costs. However, most of these robust consensus models pre-set uncertainty level parameters to control the fluctuation range of uncertain parameters, resulting in results that may be too conservative. In order to make full use of the stochastic nature of fluctuating data when establishing the uncertainty immune solution, this paper uses the chance-constrained robust optimization method to develop a robust cost consensus model. Then the appropriate confidence level can be determined to deal with uncertainty according to the actual situation, so as to guarantee the stability of the model while reducing its conservatism.
   Finally, since decision makers tend to accept the preferences of decision makers with whom they have an empathetic relationship, they are important reference information for generating preference adjustment recommendations. This paper combines empathetic evolutionary preferences with optimal adjustment preferences to generate adjustment recommendations for decision makers.
   This paper abstracts the government (the moderator) and 20 emergency experts (decision makers) from the realistic group decision-making problems in the formulation of the epidemic prevention and control plan to conduct consultations. Then a concrete implementation of the proposed framework is demonstrated, in which decision makers’ preferences and empathy relationships are randomly generated by computer. Through computer simulation experiments and comparison with other consensus models, the practicability and effectiveness of the chance-constrained robust optimization method in large-scale group consensus decision-making problems are verified. The research results show that considering that the empathy relationship can promote the group consensus, the chance-constrained robust consensus model can better balance economy and conservatism.

Multi-attribute Decision Making Method Considering Reliability of Online Reviews with Attribute Association

WANG Lanlin, LI Dengfeng

2024, 33(8):  109-114.  DOI: 10.12005/orms.2024.0258

Asbtract ( )   PDF (969KB) ( )  

References | Related Articles | Metrics

In recent years, with the rapid development of computer technology and e-commerce, the online reviews given by consumers or the public for the products can be seen on various websites, and the online reviews have a certain guiding role in the purchase decision of potential consumers. At present, consumers evaluate products or programs from multiple attributes, such as service level, logistics level, product quality, function, price and so on. Consumers use words such as “excellent, good, medium, and poor”, to express the evaluation of attributes of products or programs. As we all known, attributes are not absolutely independent of each other, and there is an association relationship between attributes actually. The online reviews on the website come from different reviewers or consumers, and the reliability of online reviews varies due to the different knowledge, experience and preferences of reviewers. Therefore, in the process of analysis and decision making, how to determine the recommendation ranking among the alternatives according to the large amount of online reviews on the website considering the differences in the reliability of online reviews with attributes association, is an important and significant research topic.
   To solve the problem of multi-attribute decision making considering the reliability of online reviews with attribute association, a multi-attribute decision making method based on the 2-tuple linguistic Choquet integral operator of online reviews is proposed. Firstly, consumers’ online reviews are mainly stored on various websites in the form of natural language. On the one hand, the use of 2-tuple linguistic to describe online reviews can effectively avoid information loss and distortion in the process of aggregation and operation. On the other hand, the deviation of linguistic is considered to make the calculation result more precise. So the 2-tuple decision matrix is constructed to describe online reviews by using 2-tuple. Secondly, the reliability of online reviews varies due to the knowledge, experience, and preferences of reviewers. In the process of evaluating customer satisfaction, if we ignore the reliability difference of online reviews and treat different online reviews equally, the accuracy of the evaluation results will be reduced. In view of this, this paper proposes a method on reliability of online reviews which considers the release time of online reviews, the number of “useful” votes obtained, the rating of reviewers, praise rewards, limited rational behavior of decision makers and so on. Thirdly, in the multi-attribute decision-making environment, it is difficult to maintain absolute independence among attributes, and the attribute association may have an impact on the decision result. Therefore, for the multi-attribute decision-making problem with attribute association, the determination of attribute weight should consider its contribution to the decision result, and the definition of Shapley value is used to determine the attribute weight in this paper. On this basis, the 2-tuple Choquet integral operator is used to aggregate the decision matrix to obtain the alternatives’ comprehensive values and ranked accordingly. Finally, in order to verify the rationality and feasibility of the proposed method in this paper, four hotels with similar orientation and room charges on the website “www.qunar.com” are taken as example, and the alternatives ranking obtained by using the proposed method is consistent with the rating ranking of each hotel on the website.
   In future work, several directions can extend our study. The online reviews in the form of graphics, videos and texts, and the influence of malicious negative reviews on decision making can be considered in the mathematical model.

E-platform Price Discount Decision with Seller Pricing and Advertising Investment

LI Li, ZHANG Hua

2024, 33(8):  115-121.  DOI: 10.12005/orms.2024.0259

Asbtract ( )   PDF (1094KB) ( )  

References | Related Articles | Metrics

In e-commerce promotion activities, the platform usually uses price discount as a promotional tool, and merchants participating in the price promotion must sell goods in accordance with the price discount released by the platform. It is difficult for low price discount to attract consumers, and the insufficient number of consumers can not attract more merchants to participate in promotional activities. Although the higher price discount attracts more consumers, because the price discount released by the platform is often borne by the merchants, the excessive price discount will affect the profits of the merchants participating in the price promotion, and reduce the willingness of the merchants to participate in the price promotion. In price promotion activities, how to determine reasonable discount pricing to attract merchants and consumers to participate in price promotion activities is an important issue for platform operation. In platform price promotion, the pricing behavior of merchants directly affects the actual purchase price of consumers, thus affecting the demand for the goods and profits of both parties. In addition, the platform profits not only come from the transaction fees of merchants, but also from the advertising investment of merchants. Therefore, the consideration of merchant pricing and advertising investment in the price discount decision of the platform has an important effect on the promotion activities.
   Based on the game theory and optimization theory, this paper studies the e-commerce market of one e-commerce platform, one merchant and many consumers from the perspective of platform discount pricing decision of price promotion activities, and builds the price promotion game model of platform and merchant. Through comparative analysis, the paper studies platform discount pricing, commodity pricing and advertising investment strategies during price promotion activities.
   We find that when the platform offers more incentives, the actual transaction price of the goods is lower, and merchants get less income. At this time, the merchants’ advertising investment is correspondingly lower. When the preferential power of the platform is reduced, the transaction price of the goods will increase, and merchants will increase the advertising investment. When the platform offers less, merchants will attract consumers by reducing prices, and the level of advertising investment will remain unchanged. Moreover, the optimal price discount coefficient of the platform is not that the lower the better, on the contrary, the optimal price discount coefficient of the platform is to maintain a high level. Although the price discount coefficient of the platform is low, it can obtain more transaction share, but the reduction of advertising investment of merchants makes the platform’s income low. When the price discount coefficient of the platform is high, merchants will find that the optimal actual transaction price on the demand curve of the commodity is lower than the price after the price discount, and merchants will achieve the optimal actual transaction price by reducing the price. In this case, the actual transaction price of merchants is high and the transaction volume is reduced, but the advertising investment of merchants increases, resulting in the final revenue of the platform.
   From the perspective of platform discount pricing of price promotion activities, this paper considers the impact of merchants’ pricing and advertising investment decisions on platform revenue, and provides theoretical support for improving the transaction efficiency of e-commerce market.

Resource Allocation Scheduling Problem with the Time-dependent Learning Effect

WANG Yichun, LYU Danyang, WANG Jibo

2024, 33(8):  122-127.  DOI: 10.12005/orms.2024.0260

Asbtract ( )   PDF (1116KB) ( )  

References | Related Articles | Metrics

In traditional scheduling model, the processing time of a job (task) is generally assumed to be a constant in advance. However, as the scheduling model evolves, this concept has been revised by introducing the concept of learning effect and resource allocation simultaneously. The importance of the learning effect is its implications in manufacturing industry and production planning, i.e., learning-by-doing increases the production efficiency. For example, a worker (i.e., machine) needs to produce a product (i.e., job), with the accumulation of experience (i.e., the learning effect), so the processing time of the product will be getting shorter. In addition, in the chemical production, the processing time of a product (for instance, a compound) can be controlled by the amount of catalyst, which is the concept of resource allocation (controllable processing time). The scheduling problems with the learning effect and resource allocation simultaneously have received widespread attention, for the research can improve production efficiency and economic benefits.
   This paper investigates a single machine resource allocation scheduling problem with the time-dependent learning effect, where the actual processing time of a job depends not only on the amount of a non-renewable resource but also on the total normal processing time of the jobs that have been processed. The goal of this problem is to determine an optimal sequence and optimal resource allocation such that the linear weighted sum of maximum completion time (i.e., makespan of all jobs) and total resource consumption is to be minimized. Under a given sequence, the optimal resource allocation can be obtained, and then the objective function (cost) of this paper can only depend on the normal processing times (i.e., this problem can be translated into a purely combinatorial optimization problem). The examples are given to show that the smallest processing time (SPT) first rule and largest processing time (LPT) first rule are not the optimal sequences. This problem is conjectured NP-hard, to solve the problem, the optimal properties, the lower bound and upper bound are given, and then the branch and bound algorithm can be proposed. An example is given to show how to solve the problem by the branch and bound algorithm. Finally, the algorithms (including the heuristic upper algorithm and the branch and bound algorithm) are tested numerically as well.
   A challenging question for future research-whether this problem is NP-hard, how to prove it, and whether there are more efficient solution algorithms for this problem. Other interesting future research might be dedicated to the extensions: either to multi-machine (for instance, flow shop or identical parallel machines) settings, or to regular (non-regular) objective functions (for instance, total weighted completion time or earliness-tardiness cost under the just-in-time (J-I-T) production environment).

Modified TOPSIS Ranking Method with Weakly Equivalent Alternatives

CHENG Youming, HU Xiangying, HE Huiyan, WU Feng

2024, 33(8):  128-134.  DOI: 10.12005/orms.2024.0261

Asbtract ( )   PDF (1244KB) ( )  

References | Related Articles | Metrics

In decision-making processes, the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) is a widely recognized multi-criteria decision analysis (MCDA) method. However, a significant challenge arises when multiple alternatives have identical ranking indices, particularly when these alternatives are located on the same multi-dimensional sphere. This leads to ambiguities in the ranking process, obscuring decision-making and reducing the clarity and utility of the TOPSIS method. Given the increasing complexity of decision-making environments in both theoretical and practical scenarios, there is a pressing need to refine the TOPSIS method to address this limitation.
   This paper aims to enhance the TOPSIS methodology by introducing a modified approach that effectively differentiates between strongly and weakly equivalent alternatives, thereby providing a more robust and clear ranking system. The proposed modification to the TOPSIS method involves several key steps to address the issue of equivalent alternatives. Firstly, this paper categorizes equivalent alternatives into strongly equivalent alternatives and weakly equivalent alternatives. Strongly equivalent alternatives are those where different alternatives have equal distances to the ideal solution, while weakly equivalent alternatives are those where the ratios of distances to the ideal solution are equal but the distances themselves are different. Using a two-dimensional target coordinate system to describe the TOPSIS system, the following characteristics are identified: (1)All alternatives are located in a specific region. (2)Strongly equivalent alternatives are at the same point, while weakly equivalent alternatives lie on a specific line. Based on the relative positions of equivalent alternatives and the positive and negative ideal solutions in the two-dimensional target space, a modified TOPSIS ranking method is proposed for evaluation systems with weakly equivalent alternatives. This method involves grouping and intergroup ranking of alternatives, and then redefining the ranking index to sort the equivalent alternatives within each group. The integration of intragroup and intergroup ranking forms a ranking sequence that achieves full ranking of systems with weakly equivalent alternatives, consistent with the classic TOPSIS method.
   The improved TOPSIS method retains the original ranking characteristics while enabling full ranking of weakly equivalent alternatives. The numerical analysis verifies the reasonableness and effectiveness of this method. The results indicate that the new method effectively addresses the ranking issue caused by weakly equivalent alternatives, enhancing decision-making accuracy and practicality. By resolving these ambiguities, the modified TOPSIS method increases the robustness and reliability of the decision-making process.
   Future research could further explore the effectiveness of this method in different application scenarios and larger datasets. Additionally, integrating this method into other MCDA techniques could lead to even more precise decision-making tools. Expanding this approach could also involve investigating its applicability in real-world decision-making situations, offering deeper insights and further validation.

Price and Service Capacity Decisions of Queueing Systems with Loss-averse Customers

CAI Xiaoli, CHEN Yao, LI Jun

2024, 33(8):  135-140.  DOI: 10.12005/orms.2024.0262

Asbtract ( )   PDF (1034KB) ( )  

References | Related Articles | Metrics

A customer’s satisfaction with a service is greatly affected by her expectations of the service attributes, with higher expectations often leading to lower utility and vice versa. Such expectations are known as reference points, and such customers are said to be reference-dependent customers. Loss aversion reflects customer behavior in the sense that compared with reference points on price and delay, losses are more painful than equal-sized gains. There is a substantial body of anecdotal and empirical evidence of customer loss aversion on price and delay attributes.Also, time is less fungible than money, and time cannot be easily saved or stored. These differences imply that people may have different degrees of loss aversion toward money and time attributes. In this study, we examine customers’ loss aversion behavior toward both price and delays in service systems where the degree of customers’ loss aversion to these attributes is different.
   We model a service system with congestion as a queueing system. Customers compare these two attributes with their rational expectations of outcomes, with losses being more painful than equal-sized gains being pleasant. Therefore, the customer’s overall expected utility from a service includes her intrinsic and gain-loss utilities. Intrinsic utility measures the direct effects of service attributes. Gain-loss utility measures the deviation of the net monetary reward and delay based on her reference points. There are different coefficients that reflect the degree of loss aversion to price and delays. We first examine customers’ equilibrium queueing strategies. We find that, unlike the traditional case in which loss aversion is not considered, there may exist three equilibrium strategies, one of which is preferred in the sense that customers’ utility is the highest at this equilibrium. Based on this, with the objective of profit maximization, we obtain three strategies on price and service capacity for different unit service capacity costs. If the unit service capacity cost is small or large, the manager will adopt strategies to attract all potential customers. However, if the unit service capacity cost is in the middle range, the manager may only attract some of the potential customers or even may not operate the system. Finally, the effect of loss aversion coefficients on the system is discussed. The results illustrate that when customers are more loss-averse to price, the optimal price and the optimal profit will increase, but when customers are more loss-averse to waiting, the optimal price and the optimal profit will decrease. However, the change in service capacity’s strategy is closely related to the unit service capacity cost.
   The management insights of this study are as follows: (1)Due to the change in the loss aversion coefficient, the threshold of unit service capacity cost will change. Managers should adopt different strategies in different threshold ranges;hence, they should pay attention to customers’ loss aversion behavior. (2)When the unit service capability cost is small or large, appropriate pricing and service capability strategies can be adopted to serve all potential customers, but when the unit service capability cost is in the middle range, appropriate strategies can be adopted to serve some potential customers, or even choose not to operate the system. (3)Managers should not ignore customers’ loss aversion towards price, however, they can ignore their loss aversion towards waiting.
   Our study will examine the reference effect on service systems when customers are loss averse toward both price and waiting time. We show that both customers’ equilibrium joining strategies, systems’ pricing and service capacity decisions are significantly different from the results without considering the reference effect. We hope our study will stimulate empirical research testing our analytical results in different service systems, such as those in the healthcare field and at call centers. Future research may also consider analyzing customers’ loss-averse behavior in observable queues.

Application Research

Research on the Impacts of Logistics Service Sharing in a Retailer-led Supply Chain

YU Yugang, GUO Dandan, ZHENG Shengming, WANG Zhaoxiang

2024, 33(8):  141-147.  DOI: 10.12005/orms.2024.0263

Asbtract ( )   PDF (962KB) ( )  

References | Related Articles | Metrics

With the rapid development of the retail industry, the retailer’s power has increased. More and more retailers have become dominant in the supply chain. In the retail industry, it has become a mainstream business model for retailers (like Amazon and JD.com) to not only resell products but also provide retail platform for the third-party sellers, fundamentally altering traditional supply chain relationships. Besides, the retailer has provided its self-built logistics service for the third-party sellers. The paper is motivated by the retailer’s logistics service sharing in practice. Our work studies the impacts of the retailer’s logistics service sharing on the common supplier, the retailer, the third-party seller, the total supply chain, and consumers in the retailer-led supply chain. This paper may be the first theoretical study on exploring the retailer’s logistics service sharing in the retailer-led supply chain and further explores the impact from various stakeholders’ perspectives, offering new insights into the dynamics of logistics service sharing.
   In this study, a game-theoretic model involving a common supplier, a retailer, and a third-party seller is developed to investigate the impacts of the retailer’s logistics service sharing. Each participant, aiming to maximize their profits, engages in strategic decision-making within the model. We first analyze the optimal pricing decisions of the supplier, retailer, and third-party seller without and with the logistics service sharing and then investigate the impacts of retailer’s logistics service sharing by comparing the game equilibria of the two cases.
   Our results indicate that the retailer’s logistics service sharing always has positive effects on the price, demand and profit of the third-party seller due to the increased logistics service level. Moreover, when the cross-service sensitivity is small, indicating that consumers are not highly sensitive to differences in logistics service quality, the logistics service sharing can lead to a “win-win-win” outcome for the retailer, the supplier, and the third-party seller. This is because the logistics service sharing can mitigate the price competition between the retailer and the third-party seller. When the service level of the retailer’s self-established logistics is not very higher than that of the third-party logistics, the logistics service sharing can improve the total profit of the supply chain. Additionally, when the cross-service sensitivity is medium, the retailer’s logistics service sharing can improve consumer surplus because the logistics service sharing enhances the service level and may result in lower retail prices. Our findings provide insights into when the retailer’s logistics service sharing is more likely to be beneficial and when it can be harmful in the retailer-led supply chain. These insights are useful for understanding the impacts of the logistics service sharing on the retailer, the supplier, the third-party seller, the supply chain, and consumers.
   There are several directions for future research. In this work, we only consider the complete information among the retailer, the common supplier, and the third-party seller. However, the retailer may have more detailed demand information than suppliers and the third-party sellers due to its rich first-hand sales data, expertise and superior forecasting ability in the selling process. Incorporating information asymmetry into the retailer’s logistics service sharing problem may be a fruitful area of research. Besides, this paper does not consider the downstream competition between retailers. In practice, retailers often engage in competition with one another. Therefore, an interesting direction for future research is to explore how the retailer’s logistics service sharing strategy might be influenced by downstream competition.

Research on Interpretable Tourism Demand Forecasting Based on JADE-TFT Model

WU Binrong, WANG Lin

2024, 33(8):  148-154.  DOI: 10.12005/orms.2024.0264

Asbtract ( )   PDF (1081KB) ( )  

References | Related Articles | Metrics

Tourism is a pillar industry in many countries, however, due to its “spatial mobility of people”, tourism is one of the industries most affected by the COVID-19 epidemic. Accurately forecasting tourism demand under the impact of the COVID-19 epidemic is very important for the strategic planning of tourism destinations and tourism-related companies. However, the uncertainty brought about by the impact of the COVID-19 epidemic has led to major challenges in forecasting tourism demand. Therefore, it is extremely important to grasp the law of changes in tourism demand as the epidemic changes, and provide guidance and advice for the tourism industry to judge the changes in tourism demand under the influence of the epidemic.
   However, most of the existing tourism demand forecasting models are limited to the selection of input variable types or the application of preprocessing methods such as synthesis or decomposition of input variables, ignoring the analysis and interpretation of the coupling relationship between tourism demand and different influencing factors. Although the existing deep learning algorithms can improve the accuracy of tourism demand forecasting, they still belong to the category of “black box” models. The lack of explanatory power creates some barriers for tourism managers to accept the research information. Therefore, it is urgent to adopt new technologies to build a new model for interpretable tourism demand forecasting under big data, and to achieve rapid decision-making and ensure forecasting accuracy in complex dynamic environments such as epidemic factors or Internet big data environments. Different from previous studies, this study introduces an explainable deep learning model for tourism demand forecasting, which can provide a comprehensive explanation for tourism demand forecasting based on multi-source heterogeneous data.
   This study proposes a novel interpretable tourism demand forecasting framework that considers the impact of the COVID-19 epidemic, by using multi-source heterogeneous data, namely historical tourism volume, new local confirmed cases, Baidu index, and weather data to predict changes in domestic tourism demand under the influence of the epidemic. This study introduces the concept of epidemic-related search engine data for tourism demand forecasting and proposes a new composition leading search index–variational mode decomposition method to process search engine data. To improve the interpretability of existing tourism demand forecasting methods, a new JADE-TFT interpretable tourism demand forecasting model is proposed, which utilizes an adaptive differential evolution algorithm with external archiving (JADE) to optimize the hyperparameters of Temporal Fusion Transformers (TFT) efficiently. TFT is an attention-based deep learning model that combines high-performance forecasting with interpretable temporal dynamics analysis, showing excellent performance in forecasting research. The TFT model produces interpretable tourism demand forecast outputs, including the importance ranking of different input variables and attention analysis at different time steps. The proposed prediction framework is verified by a real case based on Huangshan tourism data. The interpretable experimental results show that epidemic-related search engine data can better reflect tourists’ travel concerns in the post-epidemic era.
   This study aims to construct an interpretable tourism demand forecasting framework considering the impact of the epidemic. In addition, this study introduces the concept of an epidemic-related search index, thereby providing new insights into tourism forecasting under the impact of the epidemic. This study has important practical implications for managers of tourist destinations and attractions. First of all, the repeated occurrence of the epidemic has led to large fluctuations in tourism demand. It is impossible to judge the fluctuation of tourist arrivals under the impact of the epidemic by relying only on the traditional low-season and peak-season information commonly used in the past. The proposed method emphasizes the importance of the epidemic-related search index, thereby improving the accuracy of tourism demand forecasting under the impact of the epidemic. In the long run, tourism authorities can apply tourism demand forecasts to support crowd management and better preparedness against COVID-19. Secondly, tourism operators can judge the impact of the epidemic on tourism demand through search indicators related to the epidemic, rather than the number of newly confirmed local cases, because the former can better reflect tourists’ concerns. Finally, the high-frequency subsequences obtained by the CLSI-VMD method are helpful for identifying the peaks and valleys of the tourism market and provide better support for the decision-making of tourism managers.
   This study also has some limitations. First of all, tourism demand forecasting considering the impact of the epidemic is very complicated, and more influencing factors can be further considered such as the impact of government policies and restrictions on the number of tourists. Second, considering the diversity and complexity of input variables by using multi-source data, other powerful deep learning models can be used. Finally, in addition to CLSI and VMD methods, other efficient composition or decomposition methods can also be used to process search engine data effectively. In the future, we will further study the problem of interpretable tourism demand forecasting under the impact of the epidemic.

Key Opinion Elements Identification of Weibo Public Opinion Hypernetwork Based on Hypergraph

ZHU Wenbin, LI Mingda, FAN Jinyan, HU Feng

2024, 33(8):  155-161.  DOI: 10.12005/orms.2024.0265

Asbtract ( )   PDF (1058KB) ( )  

References | Related Articles | Metrics

With the rapid development of Internet technology and the popularity of social media, Weibo, as one of the largest social media platforms in China, has become an important source of online public opinion. Due to the large number of Weibo users and the diverse and complex states of public opinion, various factors are intertwined, forming an intricate system. In this system, individual users, Weibos, or comments often play a key role in the evolution of public opinion. Therefore, effectively identifying and analyzing these key elements of public opinion is of great practical significance for monitoring and managing online public opinion.
   Based on hypergraphs, this study constructs a Weibo public opinion hypernetwork analysis model, dividing the Weibo public opinion system into three subnets: social, content, and emotional. The social subnet takes Weibos as hyperedges and users who comment on the Weibos as nodes. The content subnet uses different themes as hyperedges and comments from Weibo users as nodes. The emotional subnet uses emotional intensity as nodes and emotional polarity as hyperedges. This multilayer subnet structure better characterizes the inherent structure and complex relationships of the Weibo public opinion system. To identify key public opinion elements, this study designs a series of algorithms based on hypernetwork characteristics. Firstly, the LDA topic extraction model is used to cluster Weibo comments by theme, identifying theme hyperedges in the content subnet. Secondly, SnowNLP sentiment analysis is employed to calculate the sentiment intensity of Weibo comments, constructing sentiment nodes and hyperedges in the emotional subnet. Finally, key public opinion elements in the social, content, and emotional subnets are identified based on hypernetwork characteristic indicators such as node hyperdegree, hyperedge hyperdegree, hyperedge degree, and information dissemination influence.
   The data for this study comes from real Weibo public opinion topics. By collecting Weibo content, comments, and user information under specific public opinion themes, a Weibo public opinion dataset is constructed. Using “Yuan Longping’s funeral” as a keyword, this study captures Weibo data from November 15, 2021, to November 17, 2021, including 519 valid Weibos, 5,696 comments, and 5,386 comment users. During data analysis, various analytical techniques are employed. Initially, statistical methods are used for quantitative analysis of nodes and hyperedges in the social, content, and emotional subnets, revealing the structural characteristics and distribution patterns of each subnet. Subsequently, key public opinion elements are identified by calculating indicators such as node hyperdegree, hyperedge hyperdegree, hyperedge degree, and information dissemination influence. Finally, sentiment analysis is combined to analyze and discuss the sentiment tendencies of key public opinion elements.
   By constructing a Weibo public opinion hypernetwork model based on hypergraphs, this study uncovers the complex relationships between various elements in the Weibo public opinion system. The theoretical results indicate that the model effectively characterizes the inherent structure and dynamic evolution of the Weibo public opinion system. Simultaneously, the public opinion element identification method based on hypernetwork characteristics proposed in this study demonstrates high accuracy and reliability, providing new ideas and methods for complex network analysis. In empirical research, this study applies the model to real Weibo public opinion topics and identifies six key elements of public opinion: active characters, communicators, hot Weibos, potentially popular Weibos, hot topics, and central themes. The empirical results show that these key elements of public opinion play an important role in the evolution of public opinion. Further analysis reveals that active characters and communicators have high influence and dissemination power in public opinion dissemination; hot Weibos and potentially popular Weibos have more attention and discussions; hot topics and central themes reflect the core content and public opinion trends.
   The application example of this study demonstrates how to apply the Weibo public opinion hypernetwork model based on hypergraphs to actual public opinion monitoring and management. Taking a popular event as an example, this study collects relevant Weibo content, comments, and user information to construct a Weibo public opinion dataset. Then, the constructed model is used to analyze and process the dataset, identify key elements of public opinion, and analyze their characteristics and emotional tendencies. Finally, based on the analysis results, corresponding suggestions for public opinion monitoring and management are proposed. This application example proves the effectiveness and practicality of the research method.

Online Platform’s Ex-ante/Ex-post Information Sharing Strategy Considering Joint Pricing and Quality Decision

LI Wenzhuo, LIN Qiang, LUO Xinggang, LIN Xiaogang, CHEN Danna

2024, 33(8):  162-169.  DOI: 10.12005/orms.2024.0266

Asbtract ( )   PDF (1067KB) ( )  

References | Related Articles | Metrics

With the advancement of information technology, online platforms can efficiently collect and analyze rich market data, such as customers’ browsing history and sales data, to obtain information about product demand. These data play a vital role in predicting consumers’ purchasing behavior and market demand. Online platforms can better establish a recommendation system to match products to each customer with advanced information technology and professional data analysis tools. Simultaneously, suppliers can take advantage of these data to make better operations decisions. However, suppliers must rely on platforms for demand information owing to the lack of detailed data and related analysis tools. In practice, an increasing number of consumers are concerned about product prices and product quality levels, and many consumers are willing to pay for high-quality products even if they are charged a higher price. Additionally, according to the time of obtaining product demand information, information sharing can be divided into ex-post and ex-ante information sharing. Specifically, ex-post information sharing means that the platforms share information with suppliers after obtaining real market demand information. Ex-ante information sharing means that the platforms and suppliers reach information sharing agreement in advance, and then share demand information after observing it. Inspired by the practice, considering quality decisions, is it necessary for online platforms that have massive market demand information to share the demand information with their partners? Under the reselling mode and agency mode, how do platforms make ex-post and ex-ante information sharing decisions?
   To address these issues, we develop a game-theoretic model consisting of an online platform and a supplier to investigate platform’s ex-post and ex-ante information sharing strategies under reselling mode and agency mode, respectively. Our main results suggest that: (i)The market demand uncertainty and the product quality cost coefficient are the key factors influencing the choice of ex-post information sharing by the platform. (ii)Under the reselling mode, when the quality cost coefficient is small, the platform will choose ex-ante information sharing; while under the agency mode, the platform is always willing to share demand information with supplier. (iii)In both reselling and agency mode, ex-ante information sharing is always good for the supplier. Specially, ex-ante information sharing can lead to a win-win outcome under the agency mode. However, ex-post information sharing may hurt the supplier. Moreover, the supplier with the ability to change product quality can encourage the platform to share information, which can not only better meet the needs of consumers, but also improve the profits of the platform and supplier. And then, the platform should formulate information sharing strategy according to different selling modes, product types, and the impact of different information sharing strategy on the supplier.
   Our model has several directions that merit discussion and future research. First, we implicitly assume that the online platform can perfectly forecast demand potential. In practice, the retailer might only obtain a vague signal about market demand. Thus, it would be interesting to study whether the platform should share information with the supplier at a specific accuracy level. Second, we focus on determining whether the platform would share truthful information with the supplier. Another possibility would be to investigate the impact of cheap talk on the platform’s information sharing and the retailer’s selling mode selection. Finally, we only build a supply chain consisting of one supplier and one online platform. In the future, we can build a supply chain containing multiple suppliers or multiple online platforms.

Availability Analysis of Shared Bikes Based on Bayesian Model

ZHOU Yu, ZHENG Ran, KOU Gang

2024, 33(8):  170-176.  DOI: 10.12005/orms.2024.0267

Asbtract ( )   PDF (968KB) ( )  

References | Related Articles | Metrics

Bike-sharing is a green, low-carbon and sustainable mode of transportation. It has potential benefits such as improving physical health, promoting travel safety and reducing carbon emissions. It is an important part of a sustainable public transport model. It will also be a crucial method to accomplish the strategic objective of “carbon peaking and carbon neutrality” in urban transportation systems. In the bike-sharing system, a small number of bikes are out of service every day. But unfortunately, the collection of availability information is complicated due to the maintenance management of shared bikes. Although modern technologies such as GPS, internet of things and cloud computing provide a rich information environment for the bike-sharing system, the fault and maintenance information of the bike-sharing system is still lacking. The fault feedback function is embedded in the bike-sharing system. However, users’ willingness to feedback the bike fault is not strong, and the availability of feedback information is also insufficient. The lack of fault information about shared bikes indeed increases the difficulty of quickly identifying unusable shared bikes from the perspective of system reliability. When analyzing the travel data of users in this paper, it is found that users frequently rent-out shared bikes within a short period of time, which may imply information about the availability of shared bikes. Therefore, a Bayesian model is proposed in this paper. According to the rental transaction data of users’ trips, the rent-out data is extracted, site attributes and the time information of users’ rent-out shared bikes are introduced, and a Bayesian extended model with covariates is constructed to estimate the unusable probability and number of unusable shared bikes in the site by using online transaction data. This paper verifies the effectiveness of the proposed method based on real data of Hohhot bike-sharing system database.
   In the application of the method, we first analyze the data of one day on August 1, 2017. Based on the hypothesis of KASPI et al., the prior probability applies the Bayesian extended model with covariates to obtain the unavailability probability of each shared bike and the number of unusable shared bikes in the site according to the cumulative rental and rent-out times of shared bikes. The results show that in the sites with high activity, if the shared bikes are canceled for several consecutive times, the availability of the shared bikes is low. At the same time, if there are multiple bikes with the same number of rentals in a day’s running time, the rentals in the peak period will have a higher probability of unavailability than in other periods. In addition, since the prior probability and values are both assumed values, the optimal parameter values cannot be determined. Therefore, this paper simulates different prior probabilities and values, and obtains the unusable probabilities of shared bikes under different circumstances. However, which specific parameter value has better simulation effect?Further verification needs to compare the actual number of unusable shared bikes at a specific time point with the estimated one of unusable bikes in this paper. Due to the lack of such data required by this project, this paper does not make a comparison, but only studies the unavailability probability and unavailability quantity of shared bikes under different assumed parameters.
   Based on the results obtained in this paper, the following suggestions can be put forward for the operation and management personnel of shared bikes: (1)The operation and management personnel can timely understand the availability level of shared bikes according to the unavailability probability of shared bikes, arrange maintenance tasks in a planned way, prevent the accumulation of maintenance tasks during peak periods, and make full use of human resources. (2)Small faults of shared bikes from deteriorating into bigger faults can be prevented, maintenance costs reduced and sustainable development achieved. (3)Timely maintenance of shared bikes can reduce the probability of users riding faulty bikes, reduce the risk of users travelling, improve the satisfaction of users travelling, boost users’ retention rates, and bring greater benefits to the long-term development of enterprises.
   Since the information of specific bikes in a single station in the bike-sharing system is incomplete, the number of unusable bikes estimated in this paper is a comprehensive number of unusable bikes for all stations. In further research, we will determine the optimal parameter value based on this, and accurately calculate the number of unusable bikes at each station.

Simulation Analysis of Epidemic Transmission Speed Considering Working Population and Different Protection Measures

NIU Lixia, YU Qian

2024, 33(8):  177-183.  DOI: 10.12005/orms.2024.0268

Asbtract ( )   PDF (1749KB) ( )  

References | Related Articles | Metrics

This essay mainly analyzes the trend of the epidemic, considering the influence of different populations and different protective measures on the trend of the epidemic. Firstly, the impact of different groups on the transmission of the epidemic is considered, and the transmission of the epidemic after the infection of staff and ordinary people in public places is simulated. Combined with some specific values of COVID-19, the transmission probability of the infectious disease, the number of contacts of the infectious person, the possibility of converting the latent person into the infectious person, and the total number of people in the system are determined. At the same time, the infection mechanism of the latent person and the mechanism that the cured person can be reinfected are added. On this basis, the infection rate coefficient is determined respectively according to the proportion of the number of contacts. Based on this, an improved SEIR infectious disease model is established, and the change in the number of patients is determined through simulation. Secondly, the impact of physical and chemical protection on the transmission rate of the epidemic is analyzed by comparing the transmission of the epidemic under different protective measures. Physical protection measures have good or bad effects due to different standards, materials and other protective effects. The grades of physical protection measures are divided into four levels for discussion, and the anti-epidemic effect of measures is quantified as 80%, 60%, 40%, 20%, and the control group is set up 0%. The simulation solution of the infection process is carried out respectively. In terms of chemical protection, the concept of antibody phase TS is introduced. During the antibody phase, people will not be infected by contact, and specific measures of chemical protection are taking drugs or health products, etc. After chemical protection, the antibody phase can be extended. This model is an improvement and optimization of the classic SEIR infectious disease model, and it adds the mechanism of relapse for recovered people and the mechanism of infection for latent people. Based on the difference between physical protection and chemical protection, we add a weight coefficient on the basis of the original SEIRS model, which plays a greater role in reflecting the epidemic prevention and control effect of different protective measures.
   The results show that the speed of epidemic development is reduced to different degrees after the staff takes physical protection and chemical protection measures respectively. The prevention and control effect of chemical protection measures is equivalent to the prevention and control effect of tertiary physical protection measures. Workers in public places across the country need a better physical protection than the general population in epidemic prevention and control work. The optimized infectious disease model is more in line with the actual situation of epidemic transmission and can better predict the number of possible infections in the system in the next stage. The antibody mechanism in the SEIRS model more truly reflects the impact of protective measures on the spread of the epidemic during the epidemic period, and the added incubation mechanism is more realistic than the SEIR model. In today’s era of universal vaccination, it is still necessary to pay attention to the infectiability of the novel coronavirus and combine physical protective measures,so as to prevent the spread of the epidemic.

Research on the Demand of Online Pharmaceutical Consumers under the COVID-19 Based on Text

ZHANG Li, ZHANG Zhen

2024, 33(8):  184-190.  DOI: 10.12005/orms.2024.0269

Asbtract ( )   PDF (1151KB) ( )  

References | Related Articles | Metrics

The sudden outbreak of COVID-19 has stimulated consumers’ online purchasing behavior, resulting in an explosive growth in the scale of pharmaceutical e-commerce transactions and an increasingly rich content of online comments on pharmaceutical e-commerce. The online reviews of pharmaceutical e-commerce contain a variety of information, including not only the overall star rating of consumers’ purchasing experience, but also detailed text comments, which hide consumers’ subjective feelings and consumption needs for product purchases. In order to promote the healthy development of pharmaceutical e-commerce and better meet the medication demand of consumers during the epidemic, it is urgent to carry out research on the demand of pharmaceutical online consumers under the COVID-19. From a theoretical perspective, this study focuses on online reviews of pharmaceutical e-commerce, and expands the application fields of text mining methods. From a practical perspective, this article studies the information contained in online reviews of pharmaceutical e-commerce, which can help pharmaceutical e-commerce better catch the sour spot consumer demand, timely identify problems in operating pharmaceutical e-commerce platforms, provide practical suggestions for platform operation and development, and improve consumer purchasing experience and service quality.
   This study uses the Python crawler tool to collect online comment data from a certain pharmaceutical e-commerce platform in 2019, 2020, and 2021, and captures a total of 176602 data from 17 categories of products. By processing data cleaning, word segmentation, and word frequency statistics, high-frequency words in online reviews of pharmaceutical e-commerce are extracted and displayed through word cloud maps. Then, the LDA theme model is used to further analyze the semantic relationships behind high-frequency words, in order to better understand the connections between high-frequency words. By summarizing each theme, the concerns and needs of consumers are clarified. Next, we construct a sentiment analysis model to classify emotions in online comments. The first step is to calculate the sentiment value of the text based on the Boson NLP sentiment dictionary. The second step is to train text at the word and word levels based on the BERT model beforehand. The third step is to connect the sentence vectors obtained in the first two steps and input the new sentence vectors into the SVM classifier for classification. The fourth step is to test the emotional classification performance of this model. The fifth step is to perform sentiment classification on all online comment data, including sentiment classification for individual text comments and individual topics. This study focuses on analyzing online reviews of negative emotions, as negative reviews often contain more suggestions related to products or services, which can help pharmaceutical e-commerce understand consumer sour spot and improve service levels.
   The main conclusions of this study are as follows: Firstly, consumers always pay attention to the effectiveness of medication use, logistics services, product prices, platform reliability and safety when purchasing pharmaceutical products online. Secondly, by comparing and analyzing the high-frequency words in online comments throughout this three-year epidemic, it can be found that before the COVID-19 broke out in 2019, consumers paid great attention to service attitude, commodity brand and purchase convenience. After the COVID-19 just broke out in 2020, consumers paid more attention to cold and vitamin medications, which may be because these medications help to prevent, control and cure COVID-19. In addition, the outbreak of the epidemic will affect consumers’ medication purchase decisions, and gradually cultivate consumers’ habit of purchasing medications online. In the late stage of the epidemic in 2021, consumers were more concerned about the cost-effectiveness of goods, the speed of purchase, and the quality of medication. Thirdly, consumers generally show a positive emotional attitude towards purchasing medications on pharmaceutical e-commerce platforms, with over 85% of positive comments. Fourthly,the negative online comments are mostly about medication prices,efficacy,quality, purchase of prescription, platform reliability, logistics packaging, and purchasing experience during the epidemic.
   Although this study has achieved certain research results, there are still certain limitations. For example, taking online reviews of pharmaceutical e-commerce as the research object, the amount of online review data obtained from this pharmaceutical e-commerce is sufficient but the subject is single. In addition, the emotional analysis model lacks comparative experiments to further verify the superiority of the model, and these issues can be continuously explored and improved in subsequent research.

Study on the “Internet Plus” Agricultural Supermarket Docking Strategy Considering Farmers’ Production Capital Requirement

WANG Lei, DAN Bin

2024, 33(8):  191-198.  DOI: 10.12005/orms.2024.0270

Asbtract ( )   PDF (1074KB) ( )  

References | Related Articles | Metrics

In order to better promote the docking of farmers with the market and to broaden the sales channels for agricultural products, China began to promote the “agricultural super-docking” model on a pilot basis in 2008. After more than ten years of development, this model not only effectively reduces the intermediate links in the circulation of agricultural products and reduces the cost of circulation, but also promotes the integrated development of the agricultural product supply chain, and realises a win-win situation for farmers, merchants and consumers. However, with time, the agricultural super-docking model has gradually generated new development problems. Due to China’s agricultural super docking model with “small farmers+large supermarkets” characteristics, supermarkets often defer the agricultural product payment considering that agricultural products are perishable and not easy to store. Although the deferred payment of goods can reduce the operating risks of supermarkets, it will also make farmers unable to obtain all the funds needed for the preparation for the next period of production in a timely manner, which will have an adverse impact on the interests of farmers as well as on the healthy development of the agricultural-supermarket docking model. Therefore, how to effectively solve the problem with farmers’ inability to obtain timely access to production preparation funds in the traditional agriculture-supermarket docking model is an urgent research issue.
   In this regard, based on the background of accelerating the implementation of the “Internet plus” agricultural products out of the village into the city project in China, this paper constructs a secondary supply chain of agricultural products consisting of farmers and supermarkets, and researches the “Internet plus” docking strategy between farmers and supermarkets, taking into account the needs of farmers’ production funds. On the one hand, the traditional offline supermarkets set up a network platform to dock with farmers and farmers stationed in the supermarket network platform to open a shop to sell agricultural products.On the other hand, farmers also maintain the original docking mode with supermarkets to meet the agricultural product sales demand of the supermarkets shops in the offline channels. On this basis, this paper firstly establishes the profit model of both farmers and supermarkets under the traditional docking mode to analyse the optimal decision-making and the optimal profit obtained by both farmers and supermarkets. Then, under the two scenarios of farmers taking preservation treatment and not taking preservation treatment for agricultural products, the profit model of farmers and supermarkets under the “Internet plus” docking mode is established to analyse the optimal decision-making and the optimal profit obtained by farmers and supermarkets, and the choice of “Internet plus” docking strategy by farmers and supermarkets. Further, this paper verifies through numerical examples that the platform royalty agreement based sales volume can simultaneously protect the interests of both farmers and supermarkets when they reach the “Internet plus” docking. In addition, this paper also analyses the impact of parameters such as farmers’ production capital demand parameters, demand transfer coefficients, and differences in price demand sensitivity coefficients between online and offline channels on the choice of “Internet plus” agricultural super docking stratege.
   The study finds that the platform usage fee agreement based on sales volume can help farmers meet their demand for production preparation funds and improve the interests of both farmers and supermarkets. On the basis of meeting the basic needs of farmers, when the demand for funds to prepare for the next production period is less than a certain threshold, both farmers and supermarkets will choose the “Internet plus” docking strategy of supermarkets keeping agricultural products fresh in the offline channel. Otherwise, both farmers and supermarkets will switch to the “Internet plus” docking strategy of farmers keeping agricultural products fresh in the online channel. In addition, with an increase in the influence of channel conflicts and channel competition factors such as the potential demand shifting from offline channels to online and the difference in demand price sensitivity coefficient brought by farmers’ entrance into the supermarket network platform, the above threshold for the transformation of the “Internet plus” agricultural supermarket docking strategy selection will also increase.

Trade Policy Uncertainty and Enterprise Financing Costs: From the Constraint Perspective of Financial Credit and Enterprise Investment

LI Xiangfa, ZHANG Zhe, XUE Weixian, WANG Dongling

2024, 33(8):  199-205.  DOI: 10.12005/orms.2024.0271

Asbtract ( )   PDF (1010KB) ( )  

References | Related Articles | Metrics

Trade frictions with the U.S. in recent years have exacerbated uncertainty about China’s trade policy, and China’s trade policy uncertainty index soared from 10.7 (monthly average) in 2007 to 687.6 (monthly average) in 2019. The rising uncertainty of China’s trade policy not only has a negative impact on China’s macroeconomic indicators such as economic growth, but also inhibits the investment and financing activities of microeconomic entities such as financial institutions and enterprises. With the continuation of Sino-US trade frictions, the economic impact of trade policy uncertainty has become an important frontier topic.
   With a rise in trade policy uncertainty, it is difficult for enterprises to form stable expectations for the future trade environment, and investment returns are faced with uncertainty, which prompts enterprises to change their investment behavior and affects their financing costs. So, as Sino-US trade frictions continue, what is the impact of trade policy uncertainty on the financing cost of Chinese enterprises? What is the influencing mechanism?
   For these purposes, this paper takes the financial data of A-share listed enterprises in Shanghai Stock Exchange from 1999 to 2020 as samples, constructs the panel time series model, and adopts the two-step optimal generalized estimation of Moments (GMM2S) and other methods on the basis of Wald heteroscedasticity and autocorrelation tests. This paper empirically examines the effect and mechanism of China’s trade policy uncertainty on corporate financing costs.
   The main conclusions of this paper are as follows: 1)Trade policy uncertainty has a significantly positive impact on the financing cost of firms, indicating that when firms are faced with high trade policy uncertainty, the financing cost of firms will increase, and the financing cost of firms in sensitive industries will be more sensitive to the change in trade policy uncertainty.2)Trade policy uncertainty affects the credit supply of financial institutions, exacerbates the “loan reluctance” and “loan caution” behaviors of financial institutions, which is reflected in the decline of the ratio of domestic and foreign currency credit balance of financial institutions to GDP, and ultimately affects the credit supply in the financial market and pushes up the financing cost of enterprises. 3)Trade policy uncertainty affects firms’ uncertain investment constraints, which leads to the dilemma of increasing financing costs when firms increase capital expenditure.
   This paper covers the impact of capital supply and demand factors on corporate financing costs, tests the intermediary variable of trade policy uncertainty on corporate financing costs, and opens the “black box” where trade policy uncertainty affects corporate financing costs. The conclusion of this paper confirms that the financing cost of firms is significantly affected by trade policy uncertainty, which enriches the research on how trade policy uncertainty affects micro firms. Since Sino-US trade frictions focus on different industries in China, this paper adds a dummy variable that reflects the sensitivity of industrial trade policy to verify the impact of different industries’ sensitivity to trade policy uncertainty on corporate financing costs.

Research on Loan-to-Value Ratios of Exporting Offshore/In-Transit Inventory Financing under Uncertain Price Based on Mixture CVaR Criterion

DIAO Shujie, KUANG Haibo, MENG Bin

2024, 33(8):  206-212.  DOI: 10.12005/orms.2024.0272

Asbtract ( )   PDF (1206KB) ( )  

References | Related Articles | Metrics

The exporting offshore/in-transit inventory financing is a unified credit granted by banks to shipping logistics companies. While offering logistics services such as maritime transportation of export goods, warehousing and supervision, shipping logistics companies provide pledged loans to export enterprises with capital needs. Under this model, banks, exporters and shipping logistics companies can achieve a mutual benefit and win-win situation. The key problem for shipping logistics companies to issue loans to exporters is how to determine the reasonable loan-to-value ratio according to the market value of collaterals. Under the financing mode of exporting offshore/in-transit inventory, shipping logistics companies are faced with the risk of collateral price fluctuation when they make decisions on the loan-to-value ratio, and different attitude to risk will lead to different decision-making results.
   This paper introduces the mixture CVaR criterion to formulate a decision model for exporting offshore/in-transit inventory financing and discuss the decision of loan-to-value ratios considering three different kinds of risk attitude simultaneously. Compared with the existing research, the mixture CVaR risk measurement is used to explore the bounded rationality characteristics of shipping logistics companies, which makes up for the deficiency: traditional CVaR can only reflect the single psychology of risk avoidance. In the basic model constructed above, at the end of the pledge period, when the exporter defaults, the income of shipping logistics companies is determined by the market price of collaterals at the moment. However, there are many uncertain factors in the liquidation and realization of collaterals, such as a series of judicial procedures caused by exporter default, bargaining problems, etc. The realization time of collaterals is prone to delay, which aggravates loan risks. At the same time, shipping logistics companies need to pay a certain transaction cost when clearing collaterals, and there also exists a liquidity risk of collaterals due to the massive quantity, which needs to be realized on the basis of the market price according to a certain discount. In this regard, this paper considers the influence of a liquidation time delay and liquidity risk on the realization of collaterals, establishes an extended loan-to-value ratio decision model, and discusses the impact of price risk and liquidity risk on the pledge decision of shipping logistics companies by comparing with the basic model. The extended model introduces the factors of liquidation and liquidity risks, and solves the problems with high loan-to-value ratio and loan risk caused by the neglect of liquidation stage in the existing research.
   The results show that under the same risk attitude, the loan-to-value ratio calculated by the extended model is lower than that of the basic model. The delay of collateral settlement and insufficient market liquidity aggravate the loan risk, so the loan-to-value ratio should be lowered to control the risk. The optimal loan-to-value ratio under the risk-seeking psychology is the highest, followed by the risk-neutral and risk-averse. The more drastic the fluctuation of collateral’s price, the lower the loan-to-value ratio. The loan-to-value ratio is negatively correlated with the freight rate in the shipping market. When the shipping market is depressed and the freight rate falls, the profit margin will be reduced, and the promotion of loan services could be increased.
   When the traditional shipping service is in recession, shipping logistics companies could actively seek new growth points, accelerate the pace of logistics financial business expansion, and improve the comprehensive competitiveness via exporting offshore/in-transit inventory financing service. In this process, it is particularly important to establish the whole-process management mechanism of loan risk, which should cover all aspects such as pre-loan review, in-loan monitoring and post-loan tracking, so as to realize real-time monitoring and early warning of risks and ensure the steady operation of financing business. At the same time, shipping logistics companies also need to especially strengthen the risk control in the liquidation stage, improve relevant systems and procedures, so as to effectively reduce the bad debt rate, ensure the safety of funds, and thus extend the shipping logistics service chain and expand the profit space.

Three-way Clustering Model Integrated Decomposition Ensemble Learning for Forecasting Stock Price

BAI Juncheng, SUN Bingzhen, GUO Yuqi, CHEN Youwei, GUO Jianfeng

2024, 33(8):  213-218.  DOI: 10.12005/orms.2024.0273

Asbtract ( )   PDF (1048KB) ( )  

References | Related Articles | Metrics

Accurate trend analysis and real-time price prediction are effective ways to achieve optimal investment returns. However, traditional forecasting methods face challenges in the financial markets, which are influenced by changes in the objective economic environment, investors’ expected returns, and other underlying factors. How to discover a reliable forecasting tool in uncertain environments and improve prediction accuracy is a scientific issue worthy of in-depth exploration.
   This paper introduces the idea of decomposition ensemble along with the theory of three-way decisions, and proposes a composite forecasting model based on three-way clustering. First, the Complementary Ensemble Empirical Mode Decomposition (CEEMD) method is used to decompose the original time series into several relatively stable sub-series, thereby reducing the complexity of the original time series while uncovering hidden information. Next, to address the different properties of the sub-series, sample entropy is used to measure the complexity of each sub-series, and a probabilistic rough set based on Bayesian risk decision is constructed to classify the sub-series into core, marginal, and trivial domains. Then, to avoid the lack of input information or interference from redundant information, a phase space reconstruction method is employed to determine the optimal input structures for Elman neural networks, extreme learning machines, and BP neural networks to predict the core, marginal, and trivial domains, respectively. Finally, the proposed model is applied to the prediction of ANY stock prices in the U.S. market, as well as to the prediction of important international and domestic stock indices and their constituent stocks.
   The method proposed in this paper demonstrates good predictive performance for stock prices, and its outstanding results can be attributed to the following factors: First, the CEEMD effectively uncovers hidden information in the time series. Second, three-way clustering enhances the adaptability of the forecasting method. Third, phase space reconstruction adaptively constructs the input structures of the neural networks. Theoretically, the integration of granular computing with decomposition and integration methods represents a beneficial attempt and exploration in constructing complex dynamic data forecasting decision models and methods. From the perspective of time series complexity, the construction of a three-way clustering model based on Bayesian risk decision and probabilistic rough set offers a new perspective to enrich the theory of three-way decisions. In practice, achieving accurate stock price predictions can enable investors to more effectively avoid future risks and provide scientific support and reference for practical investment decisions.


Can Listed Companies’ Interactive Information Communication Reduce Stock Mispricing?

BU Jun, SUN Guangguo, XU Junjie

2024, 33(8):  219-225.  DOI: 10.12005/orms.2024.0274

Asbtract ( )   PDF (967KB) ( )  

References | Related Articles | Metrics

One of the foundations for the optimal allocation of capital market resources is that the stock market price can reasonably reflect the market value of market participants, particularly listed companies. The long-term deviation of the market value reflected by the stock prices of listed companies from their intrinsic value, i.e., stock mispricing, not only poses a serious risk management challenge to listed companies, investors and regulators, but also leads to inefficient investment at the micro-firm level and hinders the resource allocation function of the securities market at the macro level. Hence, whether and how stock mispricing can be improved through the disclosure, communication, and regulation of company information has long been a topic of concern for academics and market practitioners.
   The capital market is essentially an information market, and information is the most important factor in determining the stock value of listed companies. Fundamentally, information asymmetry and uneven distribution of information lead to long-term deviations of listed companies’ stock prices or market value from their intrinsic value, resulting in stock mispricing. With the utilization of information technology and the improvement of market maturity, the information transmission mode between listed companies and investors is changing from “one-way information disclosure” to “interactive information communication”. In this context, the SZSE and SSE established the interactive information communication platforms of “SZSE Interactive Easy” and “SSE E-Interaction” in 2010 and 2013 respectively. In recent years, scholars have also begun to pay attention to the impact of the establishment of interactive communication platforms in China’s capital market on investors, firms and the market. However, existing literature has rarely addressed the impact of behavioral characteristics such as the timeliness of responses and response rates in interactive communication by listed companies, nor has it directly examined the impact of interactive communication behavior on stock market mispricing.
   In this context, this paper utilizes data from non-financial A-share listed companies from 2014 to 2019 to examine the relationship between companies’ interactive information communication behavior and its stock mispricing. In measuring companies’ interactive communication behavior, this paper uses firms’ response frequency and response timeliness in interactive Q&A with investors to measure firms’ positive interactive behaviors. For measuring stock mispricing, this paper references the work of RHODES-KROPF et al. (2004), using the market-to-book ratio decomposition method to derive and calculate stock mispricing. After controlling a series of influencing factors, this paper constructs a regression model to test the main research relationships. In addition, this paper also uses the Sobel mediation test proposed by BARON and KENNY (1986) to test the mechanism effect by modeling the equations and estimating the significance level of the Sobel Z value. Furthermore, in the robustness test, the paper employs PSM, Heckman’s two-stage and DID methods to mitigate the endogeneity problem and derives more robust findings by altering the measurement of key indicators, controlling stock pricing efficiency at the industry level, and comparing different China’s Exchanges.
   The results of this paper indicate that the more active the behavior of listed companies in interactive information communication with investors, i.e., the higher the interaction frequency and timeliness of interaction, the lower the degree of deviation of the company’s stock market price from its intrinsic value, and the lower the degree of stock mispricing. The mechanism of the effect shows that company’s positive interactive communication behaviors reduce the degree of stock mispricing by reducing heterogeneous investor beliefs, increasing the information content of the stock price and optimizing the company’s information environment. In addition, given that market participants in interactive platforms are predominantly small and medium-sized investors, further research also finds that the positive governance effect of company-investor interactive communication on stock mispricing will be more prominent when investor sentiment is stronger and the proportion of institutional investors’ shareholding is smaller. The findings of this paper confirm the market effectiveness of interactive communication platforms under the supervision of China’s Exchanges, and also provide new evidence for optimizing the means to further improve the efficiency of market resource allocation.

Management Science

Research on Volatility Spillover between Real Estate Stock Price and High Technology Sector Based on DGC-t-MSV Model

XU Ye, XIE Tao, TAO Changqi

2024, 33(8):  226-232.  DOI: 10.12005/orms.2024.0275

Asbtract ( )   PDF (1165KB) ( )  

References | Related Articles | Metrics

With the wide opening of the financial market, the internal connection between the traditional industry and emerging one is gradually strengthened. Since September in 2021, there have been frequent debt defaults of housing enterprises, some stocks have fallen by more than 50% since the peak of 2018, the Hang Seng Real Estate Construction Index has fallen by nearly 18%, and stocks such as China Evergrande, China Fortune Happiness and Sun City Group have fallen by more than 50% this year. Short bursts and shocks have become the basic trend of the real estate sector, and it is more necessary for the high-tech sector to assume the important task of guiding the flow of social funds and dispersing price risks in the stock market. Exploring the risk linkage structure and spillover effect mechanism between “the real estate and high-tech” has important academic value and practical significance.
   Existing researches, to some extent, have ignored the actual impact of real estate stock price fluctuations on the high-tech sector in the stock market. To this end, based on Maximum Overlapping Discrete Wavelet (MODWT) decomposition and DGC-t-MSV model, this paper focuses on the spillover effect of real estate stocks on high-tech sectors. First, wavelet theory is widely applied to the multi-dimensional decomposition of financial time series. MODWT is not limited by ordinary discrete wavelet for the amount of data, also known as displacement invariant discrete wavelet transform or stationary wavelet transform, and after each decomposition of the same amount of data, has stronger applicability than ordinary discrete wavelet transform. Second, stochastic volatility model can better describe the real characteristics of financial volatility by introducing stochastic process. The DGC-t-MSV model constructed in this paper can estimate the dynamic correlation coefficient of the return series and obtain the intensity and direction of spillover effect based on Granger causality test.
   Based on the index compilation methods published by China Securities Index Company and Shenzhen Securities Information Company Limited, this paper extracts the real estate blue chip stocks in Shanghai and Shenzhen stock markets to build the real estate top50 index (RE). First, the total average daily market value and average daily turnover of the selected stocks in the sample space in the last half a year are collected, and the stocks that rank the bottom 20% in average daily turnover are excluded. Secondly, the top 50 stocks are selected to form the initial sample stocks according to the total average daily market value from the highest to the lowest. Finally, the real-time price index is calculated according to the Pai weighting method, and the base date of the index is April 20, 2003. The index base point is at 1000 points. The construction method of high-tech industrial Composite index (HT) is the same as above.
   In this paper, WinBUGS software is used for MCMC iteration. The results show that: (1)There is a dynamic correlation between the real estate and high-tech sector, and volatility spillovers in real estate stocks are stronger. (2)The restrictions on innovation inputs brought by the real estate industry are gradually reduced after the government’s restrictive policies. (3)Frequent trading in the short and medium term is the main reason for the real estate premium, and the high-tech sector needs to be alert to the spillover risks faced by the high-tech sector. Therefore, actively guiding value investment in the high-tech sector and reasonably allocating technology elements can promote the high-tech industry development effectively.

Service Quality Disclosure Strategy of Service-oriented Enterprises Based on Customer Uncertainty Decision

JIANG Tao, GAO Li, JIANG Tao, CHAI Xudong

2024, 33(8):  233-239.  DOI: 10.12005/orms.2024.0276

Asbtract ( )   PDF (1277KB) ( )  

References | Related Articles | Metrics

In the service industry, service quality is an important factor in determining customers’ willingness to pay and enterprises’ profits. However, due to the intangible nature of service products, the service quality information held by enterprises and customers is asymmetrical, for enterprises can obtain accurate service quality information through customer feedback, survey, expert evaluation, market data and other ways, while customers lack resources and channels to obtain such information, unless enterprises disclose service quality information. When the enterprise does not disclose the relevant service quality information and customers face uncertain service quality, different customer groups usually have different psychological characteristics and adopt different decision-making schemes according to their own preferences. In view of the above background, this paper focuses on the following issues. When enterprises hide the information of service quality, how can customers make decisions with uncertain decision criteria and how can enterprises reasonably set service price? How do enterprises develop service quality disclosure strategies to maximize service revenue?
   In order to study how the uncertain decision made by customers affects the choice of service quality information disclosure strategy in the face of the uncertainty of service quality, this paper takes an M/G/1 queue as the background, by establishing the customer service utility function and the enterprise service revenue function under the two strategies of whether the service quality information is disclosed or not. We first assume that when enterprises disclose the service quality information, customers can obtain service quality information before making decisions, and there is no difference among all customers. When enterprises hide the service quality information, there are two types of customers in the queue: optimistic customers and pessimistic customers. In the face of uncertain service quality, optimistic customers adopt the risk criterion, while pessimistic customers adopt the conservative criterion, where the decision maker who applies the risk criterion (conservative criterion) must first determine the maximum (small) return value that each alternative scheme may lead to, and then compare the maximum (small) return value of these alternative schemes. Then the largest one is selected, and the alternative corresponding to this maximum value is the decision maker’s final choice. Based on the above assumptions, by using the queuing-game theory, the paper analyzes the related influencing factors leading to the change in the equilibrium strategies of customers and service pricing decisions of enterprises, and discusses the impact of service quality information disclosure on the enterprises’ revenue.
   The results show that, first, if pessimistic customers adopt the “all balk” strategy, optimistic customers will adopt either the “all join” strategy or a “mixed strategy”. If pessimistic customers adopt a “mixed strategy”, optimistic customers will adopt the “all join” strategy. Under the two strategies of whether the service quality information is disclosed or not, the enterprise throughput decreases with an increase in service price. In addition, when the service price is small, the disclosure of service quality information can improve the enterprise throughput, and vice versa. Second, if the service quality and the proportion of optimistic customers in the arriving population increase, the strategy of disclosing service quality information cannot always obtain a large optimal service price and service revenue. When the service quality information is hidden, the optimal service price is discontinuous in terms of the proportion of optimistic customers in the arriving population, but there is a jump point. When the proportion of optimistic customers is greater than the jump point, the optimal service price tends to decrease. Third, when the proportion of optimistic customers in the arriving population is relatively small, enterprises should always adopt the strategy of disclosing service quality information. As long as the proportion of optimistic customers in the arriving population exceeds a certain value, higher service revenue can be obtained by choosing to hide service quality information.