Abstract: After emergency, the first step of emergency response is to establish emergency rescue network, and allocate reasonable emergency resource to assure relief process going on smoothly. In order to improve efficiency of rescue, maxize supply and minimize loss of disaster, emergency relief network plan is to setup some temporary relief centers and allocate different types and quantities of emergency resourcs. This paper proposes a second-stage mixed integer programming model for determining relief network plan for hurricane disasters. Demand information is random and represented by intervals. The proposed model is solved using robust optimization, which makes use of interval structure of uncertain demand information. Robust counterpart of the model is more conservative and it can't reflect the effect of the second stage decision variables. We introduce robust counterpart of the model and solve it by limiting the second stage decision variable as linear function of uncertain parameters. Finally, we propose a case study to highlight efficiency of the proposed solution strategy

Key words: emergency management, emergency rescue network plan, uncertain emand, robust optimization

摘要： 突发事件发生后,快速应急响应的第一步是启动应急救灾网络,合理配置应急救灾资源,以保证救灾过程顺利进行,提高救援效率。本文以台风灾害为背景,建立二阶段应急救灾网络的混合整数规划模型,解决台风灾害的灾后应急救灾网络的规划与设计问题。在求解模型时,考虑需求信息的分布难以确定,并且在一定范围内变动的特点,利用鲁棒优化的方法处理不确定性需求,从而得到合理的临时救灾中心选址以及应急资源配置信息。数值试验表明,建立的模型是实际可行的,而且算法也是有效的。

关键词: 应急管理, 应急救灾网络规划, 不确定需求, 鲁棒优化