Abstract: The large-scale disasters have seriously destroyed or collapsed the original social and economic systems in the affected areas. After the disaster, the currency could not circulate, the infrastructure is damaged and the supply chain of economic market is disrupted. These problems often lead to the shortage of emergency supplies. Furthermore, the lack of basic survival and living materials will lead to the discomfort of individuals and bring anxiety and pain to the affected people. To reduce the human suffering and conduct the relief operation effectively, it is urgent for decision makers to incorporate beneficiaries' perception into total cost of emergency relief. In this paper, a numerical rating scale (NRS) is designed to construct a human suffering function, which can describe the suffering perception cost of the beneficiaries.According to the principles in welfare economics, the concept of social cost is first proposed in emergency management. It refers to the sum of logistics cost and suffering perception cost. An integrated optimization model is proposed to combine evacuation and resettlement of the affected people, location of temporary shelters and distribution of relief materials, whose objective is to minimize total social cost. The classical mixed integer programming method and improved genetic algorithm are designed to solve the model effectively. Finally, the performance of the two methods is compared by numerical simulation.Comparisons of three simulation scenarios indicate that the number of transferrable population obtained by using genetic algorithm is larger and the total social cost is smaller as the increase of the maximum rescue time. The model and algorithm are applied to the case of 2014 typhoon rammasun in Hainan province in China. The results further show that in order to maximize the rescue effectiveness, more shelters need to be built and sites should be located near the affected areas.In addition, each shelter needs more transport vehicles to evacuate more people in a shorter rescue time. It provides an effective suggestion to help post-disaster managers to enhance the effectiveness of emergency rescue operations.

Key words: emergency management, human suffering, emergency shelters, location-allocation, improved genetic algorithm