Integrated Optimization Of Depot Selection And Vehicle Routing For Relief Resource Distribution In Urban Public Emergencies

With the rapid development of social economy and the acceleration of urbanization process, a range of urban public emergencies that cause huge casualties and economic losses occur frequently. Urban public emergency response is a complex and systematic project, one of the key elements is to ensure timely supply of all emergency resources. Urban public emergency logistics system must meet the logistics needs under constraints of space, time and resources in order to maximize the time effectiveness and minimize disaster losses.Depot selection and vehicle routing are two very important problems in the emergency resources distribution. Presently they were studied separately in the most of previous literatures. However, they are so closely related that there is a need to study jointly. In this paper, depot selection-routing problem is discussed with the background of urban public emergencies.Firstly, a travel time calculation method of roads was proposed considering cross-time traveling and first in first out principle in the real-time and time-dependent networks. It provided a method to compute the model objective function of the main problem.Secondly, a fuzzy dynamic DSRP in the case of single-accident point and multi-rescue points was studied. We developed a mathematical model of DSRP to minimize response time of emergency, considering some constraints, such as multiple vehicles, facility capacity, fuzzy demand and so on. Fixed decision-making cycle was applied to make dynamic decision, a two-stage heuristic algorithm integrating improved genetic algorithm and linear programming was proposed to solve the problem, then the validity of the model and algorithm were illustrated by numerical example.Thirdly, a fuzzy dynamic DSRP in the case of multi-accident points and multi-rescue points was studied. Based on the research of single-accident point, we increased the rescue time window constraint, aimed to maximize response time satisfaction. For real-time traffic conditions and new accident points, a strategy combining event triggered decision and fixed decision-making cycle was adopted to make dynamic decision. We followed the algorithm in previous chapter, and illustrated the scalability of the algorithm by numerical example.