Abstract: In recent years, with the frequent occurrence of various types of emergencies such as COVID-19, fire, earthquake and so on, how to carry out emergency rescue effectively and efficiently has become a crucial research topic. Emergency rescue, however, is highly uncertain and dynamic due to complex environment and urgent resource deployment. According to a statistic data analysis, the main reason of the accident is that there is no flexible contingency plan to guide the rescue to go on in order and along with serious shortage of rescue materials and professional person. Existing contingency plans for emergencies have only concerned the establishment of the rescue organization, division of personal responsibility, but they can notgive specific guidance for resource and time scheduling. Therefore, developing a robust schedule before the accident plays a great role for emergency rescue. A stable and reliable rescue plan and its effective response can greatly reduce the risk of disaster diffusion and serious economic losses. By using the theory of project scheduling under uncertain environment, this paper studies the emergency rescue problem with random resources interruption. On the one hand, before the accident occurs, the proactive project scheduling method is used to develop a high robust baseline schedule to guide the rescue process to go on in order. On the other hand, once an accident occurs, the reactive project scheduling method is invoked to take appropriate strategies in response to disruptions in order to ensure the feasibility of the baseline schedule as far as possible and also the smooth development of the rescue work.
Based on this background, an emergency rescue scheduling optimization model is formulated with two sub-models included. The first sub-model tries to maximize total project robustness where both time buffer robustness and resource buffer robustness are considered under the constraint of renewable resources and project deadlines. The second sub-model indeed tries to minimize the adjustment losses once disruption occurs. The two sub-models are associated with the baseline schedule that the former provides an original baseline for determining the actual start time of activities, and the latter will adjust the baseline with the help of simulation tool at each disruption point. Through the integrated optimization of the above two sub-models, we can obtain the quantitative coordination between the preparation of emergency rescue plan and emergency disposal, and ensure the most effectively realization of the rescue objectives.
For the NP-hardness of the problem, an improved tabu search heuristic algorithm is developed where the solution is represented by an array of four lists, including an execution mode list, an activity list, a time buffer list and a resource buffer list. At the same time, in order to further improve the efficiency of the algorithm search, the algorithm neighbourhood search strategy is developed based on the characteristics of the solution representation. Then, the model and algorithm are applied to the case of a real blowout event in DG oilfield. In order to highlight the urgency and importance of emergency rescue, the accident case is mainly divided into three stages, and according to the changes of the key process data in the development of the accident, three stages are simulated respectively, and then the simulation results of different stages are analysed in detail. Moreover, the performance of the robustness baseline is compared with the actual schedule of the minimizing project makespan. The results show that the robust scheduling model can be used to arrange the schedule plan in emergency rescue, so that the activities can obtain appropriate and reasonable buffer and better resist the uncertain factors. Furthermore, when the baseline schedule has to be changed due to unexpected circumstances, the reactive scheduling research results can be used to adjust the baseline quickly and effectively, decreasing the total adjustment loss by about 20.4% compared with the makespan minimization model. It can be seen that through the cooperation of proactive and reactive scheduling methods applied in the emergence rescue, it can effectively improve the baseline robustness and strengthen its capability against resource disruption case, and moreover reduce the project loss to a greater extent. The conclusion of this paper can provide constructive adviceand flexible guidance for similar accident emergency rescue work. Although this paper has done some innovative work on the robust scheduling problem of emergency rescue projects, there is still a lot of problems that need to be solved. In the future, we should consider the scheduling problem of emergency rescue activities with preemptive rights. In the actual emergency rescue process, there are priorities among various activities, so it is also of great significance to consider the right of preemption.

Key words: emergence rescue, proactive/reactive project scheduling, time/resource buffer, blowout event

摘要： 突发事件应急救援具有高度的不确定性与动态性,稳定可靠的救援计划及其合理有效的应急处置则可以大幅度降低灾害蔓延的风险。通过运用不确定环境下项目调度的理论与技术对资源随机中断情形下的突发事件应急救援问题展开研究,一方面在事故发生前,采用前摄性项目调度方法制定鲁棒性水平较高的基准计划指导救援过程有序开展;另一方面在事故发生后,借助反应性项目调度方法随时针对环境变化采取恰当的处置策略做出快速有效的应急响应,减小事故进一步扩散的风险。研究结果表明通过前摄性与反应性调度方法的协作与配合,可有效应对突发事件应急救援中发生的资源中断情况,更大限度地减少项目的损失。

关键词: 突发事件应急救援, 前摄性/反应性项目调度, 时间/资源缓冲, 井喷事故