我国是一个地质灾害多发的国家，以地震、巨型滑坡为代表的重大地质灾害频繁发生，严重威胁着人民的生命安全。重大地质灾害事件由于其突发性、强大破坏性、时间紧迫性、高度不确定性以及衍生和演化等特点，成为应急管理领域研究的重点和难点。目前急需提升对这类灾害的应急响应能力，加强应急救援和响应工作。 为了更有效地应对复杂性突发事件，2018年初我国成立了应急管理部，提升了应急管理的统筹协调水平。由此也带来了一系列挑战，面临着应急资源的整体统筹优化，大量不同类型应急救援队伍之间的融合等问题。对此，本文以重大地质灾害应急救援为背景，研究了灾害救援过程中两个阶段的问题。 第一个是救援队伍的区域性部署决策问题，属于应急资源的整体统筹优化范畴。当应急救援队伍到达现场以后，快速有效的实现救援队伍的区域性部署是应急救援工作开展的前提和基础，同时也是控制灾情进一步恶化、减少受灾点人员伤亡的重要保障。 第二个是受灾区域内救援队伍间的合作救援问题，属于应急救援队伍之间的融合范畴。在完成救援队伍部署工作后，救援队伍将被分配至相应的受灾区域，灾害现场搜救行动处于一个快速多变的环境，系统状态瞬息万变，为保证应急决策的时效性，需要救援队伍具有一定的“自主性”，能有效的进行自主决策，确定“是否合作?跟谁合作?如何合作?”。针对此，借鉴了分布式任务分配模型对救援队伍间的合作机制进行设计，并考虑了两种不同的队伍编制情况。本文主要的研究工作如下： (1)救援队伍的区域性部署策略研究。重大地质灾害发生以后，影响范围广，产生了多个受灾区域，然而参与救援的人力资源极其有限，多个受灾区域对救援队伍的需求难以同时被满足，因此需要综合考虑多种因素将有限的救援队伍公平、有效地部署至各个受灾区域。需要考虑的因素主要包括三方面，即救援优先级，灾民满意度以及救援队伍分配效用。对此，本文提出了基于救援优先级的救援队伍区域部署决策方案。首先，对受灾区域进行分析，确定各个区域的受灾状况。然后，对各个受灾区域的救援优先级进行评估。最后，以受灾区域的灾民满意度最大化和救援队伍分配效用最大化为目标，建立了救援力量的优化分配模型，并采用NSGA-II算法对该多目标优化模型进行求解，为灾害现场救援队伍区域性部署决策提供辅助。 (2)同一编制的救援队伍合作搜救机制设计。同一编制指的是每支救援队伍构成类似，并且每一支救援队伍可独立完成对一个伤员掩埋点的救援，不过可能会消耗太多时间，导致贻误最佳救援时间。针对同一编制的救援队伍之间的合作搜救过程，设计了救援队伍间的基于分布式拍卖算法的合作救援机制，通过多主体仿真(Agent-basedModeling and Simulation，ABMS)和蒙特卡洛方法(Monte Carlo method)对提出的合作搜救方案进行评估，并与其它方案进行比较研究，包括非合作救援与基于F-Max-Sum算法的合作救援方案，其中非合作救援指的是现行的缺乏有效合作机制的灾害救援行动。基于该合作机制的救援方案明显优于非合作救援，并且具有统计意义上的显著性；与知名的F-Max-Sum算法相比，性能并无逊色，并且计算过程复杂度低，更适用于灾害环境中的快速决策。 (3)不同编制的救援队伍合作搜救机制设计。同样是动态灾害环境中救援队伍间的合作救援问题，但是该部分研究内容考虑了不同编制救援队伍间的合作问题。该部分对救援队伍的编制情况做出了不一样的假设，即假设救援队伍构成的差异性很大，每支救援队伍所具有的救援能力不同。完成每一个救援任务所需要的能力也不同，需要根据该项救援任务的具体特点进行评估。基于此，该部分内容主要针对灾害现场不同编制救援队伍之间的合作救援，设计了基于组合拍卖机制的合作救援方案，并与其它救援方案进行了比较分析。首先，比较分析的结果显示在三种不同情景下，基于组合拍卖的合作救援机制的仿真结果指标均在0.01的显著性水平下优于非合作救援，与基于F-Max-Sum算法的合作救援方案对比，除了在计算过程的复杂程度上优于后者，仿真结果指标也在一定程度上略优于后者。其次，救援队伍的数量对救援效率的影响分析结果显示救援数量增加对救援效果的改善存在边际递减效应。最后，对提出的合作救援方案在极端情况出现时的可靠性进行了分析，即鲁棒性分析。 (4)区域部署策略与合作搜救过程的集成研究。将救援队伍的区域部署策略问题与灾害现场救援队伍间的合作搜救过程集成到一个实例分析中进行了研究。研究了三种不同部署策略对合作救援仿真结果的影响，第一种是考虑区域救援优先级与救援队伍-受灾区域匹配度的策略，第二种是不考虑区域救援优先级的救援队伍分配模型，第三种是不考虑救援队伍-受灾区域匹配度的救援队伍分配模型。分析结果表明考虑了区域救援优先级与救援队伍-受灾区域匹配度的策略，在救援效果方面，优于不考虑区域救援优先级的策略，和不考虑救援队伍-受灾区域匹配度的策略。不考虑救援队伍-受灾区域匹配度的策略使得所有救援队伍未能发挥出自己的特点，造成了资源浪费，对整个受灾区域的救援效果造成了最严重的影响。 本文的研究工作主要是关于重大地质灾害后救援队伍资源的统筹优化与救援队伍间的合作机制设计，创新点有：(1)以受灾区域人口的需求满意度最大化和救援队伍分配效用最大化为目标，建立了救援力量的优化分配模型，提出了基于救援优先级的救援队伍区域部署决策方案；(2)针对灾害现场救援过程中救援队伍之间缺乏有效合作的现状，考虑了两种不同的队伍编制情况，并分别结合不同的救援队伍编制特点，设计了于分布式拍卖的合作救援机制与基于组合拍卖的合作救援机制；(3)传统研究中往往只单独研究了救援队伍的区域性部署问题，本文将救援队伍的部署决策与受灾区域内救援队伍间的合作搜救过程集成进行了研究。 关键词：救援队伍；区域性部署；合作搜救机制；拍卖

 The geological disasters occur frequently in China, such as earthquakes, giant landslides, which have caused mass casualty incident. Large-scale geological disasters become the focus and difficulty in the research of emergency management, due to their suddenness, destructiveness, time urgency, high uncertainty, derivative and evolvement. Therefore, it is urgent that we improve the emergency response capability for such large -scale disasters. In order to cope with the serious and complex emergencies efficiently, the Ministry of Emergency Management of the People's Republic of China was established in early 2018, which has improved the overall coordination for emergency rescources in emergency management. This also brings a series of challenges, such as the optimized allocation of emergency resources, the integration of a large number of different types of emergency rescue teams and so on. In this regard, two main parts in two stages of the large-scale geological disaster rescue process are studied in this paper. The first part is about the regional deployment of the rescue teams, which belongs in the category of emergency resources allocation. When the emergency rescue team arrives at the disaster scene, the rapid and effective realization of the regional deployment of the rescue teams is the premise and basis for carrying out the emergency rescue operations efficiently. It is also an important guarantee to stop the deterioration of disaster situation and reduce the casualties and property losses of the affected people. The second part is about the rescue cooperation between the rescue teams in the disaster-stricken areas, which belongs to the category of integration of emergency rescue teams. After the regional deployment of rescue teams, the rescue team will be assigned to the corresponding disaster-stricken area. The search and rescue is carried out in a fast-changing environment, and the system state changes rapidly. To ensure the timeliness of emergency decision-making, the rescue team needs to make decisions on their own, and determine "whether to cooperate? Who to cooperate with? How to cooperate?" In this regard, the task allocation model in Multi-robot system is used for the mechanism design of cooperation between rescue teams. Two researches are conducted here according to the characteristics of rescue teams, i.e. homogeneous rescue teams and heterogeneous rescue teams. In the first case, the rescue teams are homogeneous, which means all the rescue teams are alomost the same in the formation, and each rescue team can complete a task seperately. In the second case, the rescue teams are heterogeneous, which means the rescue teams are highly specialized, and each of them has its own capability. Thus, one rescue team often can not complete a task separately. Only when two or more rescue teams work as a coalition, can the task be completed. The main contents and conclusions of this paper are as follows: (1)Research on the regional deployment of the rescue teams. The occurrence of large-scale geological disasters often results in several disaster-stricken areas for its strong influence. However, the human resources involved in the rescue are extremely limited. Each disaster-stricken area's demand for the rescue teams is impossible to be satisfied at the same time. Therefore, it is necessary to consider a variety of factors in order to deploy the rescue teams to all disaster-stricken areas fairly and efficiently. In response to the multiple competing demands of disaster-stricken areas, this paper proposes a regional deployment strategy of rescue teams based on rescue priority. First, the disaster status of each disaster-stricken area is captured through field investigation. Second, the rescue priority of each disaster-stricken area is evaluated. Finally, aiming at maximizing the satisfaction of disaster victims in the disaster-stricken areas and maximizing the utility of allocation, the optimal allocation model of rescue teams is established. The multi-objective optimization model is solved by NSGA-II algorithm, thus it can provide assistance for regional deployment of rescue teams. (2)The simulation optimization of rescue cooperation between homogeneous rescue teams in disaster relief. The rescue cooperation between homogeneous rescue teams is actually about task allocation in a dynamical environment. The rescue team needs to make decisions based on their knowledge, in order to form a coalition efficiently. We have developed a cooperative rescue plan for the homogeneous rescue teams based on distributed auction mechanism, and evaluated the proposed rescue plan through agent-based modeling and simulation (ABMS) and Monte Carlo method. In order to understand the performance of the proposed rescue plan, we compare it with non-cooperative rescue plan and F-Max-Sum-based rescue plan. The result shows that the proposed cooperative rescue plan could improve the rescue efficiency significantly. Compared with the non-cooperative rescue plan in the three scenarios, it increase victims' relative survival probability by 7%-15%, increase the ratio of survivors getting rescued by 5.3%-12.9%, and decrease the average elapsed time for one site getting rescued by 16.6%-21.6%. Furthermore, it performs somewhat better than the well-known F-Max-Sum-based approach in regard to some indicators. Search and rescue programs are evaluated and compared with other programs. Particularly, the low complexity of computation has made our proposed cooperative rescue plan more appropriate for the cooperation among rescue teams in disaster relief than F-Max-Sum-based approach. The robustness analysis shows that search radius can affect the rescue efficiency significantly, while the scope of cooperation has little effect on the rescue efficiency. The sensitivity analysis shows that the two parameters, the time limit for completing rescue operations in one buried site and the maximum turning angle for next step, both have great influence on rescue efficiency, and there exist optimal value for both of them in view of rescue efficiency. (3)The simulation optimization of rescue cooperation between heterogeneous rescue teams in disaster relief. In this case, the tasks are heterogeneous so that the tasks request different kinds of capabilities to complete, and the request could be satisfied by a coalition which is comprised of two or more rescue teams. The rescue teams are also heterogeneous for that they possessed different kinds of capabilities. Therefore, we have proposed a cooperative rescue plan for heterogeneous rescue teams based on a combinatorial auction-based task allocation scheme, and evaluated it through agent-based modeling and simulation (ABMS) and Monte Carlo method. Compared with the non-cooperative rescue plan in the three scenarios, it increase victims’ relative survival probability by 13.8%-16.3%, increase the ratio of survivors getting rescued by 10.7%-12.7%, and decrease the average elapsed time for one site getting rescued by 19.0%-26.6%. Besides, the proposed cooperative rescue plan has outperformed the rescue plan based on F-Max-Sum a little bit. Moreover, we have presented an analysis to show how the number of rescue teams influences the rescue efficiency. The result shows that the increase in the number of rescue teams has a marginal diminishing effect on the improvement of rescue efficiency. Finally we have presented the robustness analysis to investigate the reliability of our proposed cooperative rescue plan under extreme operative situations. The robustness analysis shows that the proposed rescue plan is relatively reliable on condition that both the search radius and scope of cooperation are larger than thresholds. (4)The integrated research of regional deployment strategies and cooperative rescue operations. We integrated the regional deployment strategies of rescue teams with the cooperative rescue operations of rescue teams in a case study. The effects of three different deployment strategies on the simulation results of cooperative rescue are studied. The first strategy has taken the regional rescue priority and matching degree between rescue teams and disaster-stricken areas into account. The second strategy has not taken the regional rescue priority into account. The third strategy has not taken the matching degree between rescue teams and disaster-stricken areas into account. The analysis results show that the strategy which has taken both the regional rescue priority and matching degree between rescue teams and disaster-stricken areas into account outperforms the other two strategies obviously. Moreover, the strategy which has not taken the matching degree between rescue teams and disaster-stricken areas into account has failed to make the rescue teams operate efficiently, thus has caused a waste of human rescources, and has created the most serious impact on the rescue efficieny of the entire search and rescue. The innovation of this paper is mainly reflected in the following three aspects: (1)We have developed a regional deployment strategy of rescue teams based on rescue priority, which is aimed at maximizing the satisfaction of disaster victims in the disaster-stricken areas and maximizing the utility of allocation. (2)Given the lack of efficient cooperative rescue plan for rescue teams in disaster relief, we have developed copperative rescue plan for the rescue teams. And we have taken the structure of rescue teams into consideration, thus we have developed cooperative plan for the homogeneous rescue teams and heterogeneous rescue teams based on auction mechanism and combinatorial auction mechanism, rescpectively. (3)In the traditional researches, the regional deployment of rescue teams is often studied separately. In this paper, we have integrated the regional deployment strategies of rescue teams with the cooperative rescue operations of rescue teams in a case study. Thus we can evaluate the regional deployment strategy through the simulation results of cooperative rescue operations, which is a novel way to evaluate the regional deployment strategy. Key Words: rescue teams; regional deployment; mechanism design of rescue cooperation; auciton