Abstract: It has very important significance to research on the location theory for emergency material depositories in response to frequently occurring large-scale emergencies. In allusion to the characteristics of tremendous and continued demands for emergency supplies during large-scale emergency occurring, considering the factors of fairness, efficiency and cost for facility location, a multi-objective multiple gradual coverage location model is proposed based on the idea of multiple coverage and gradual coverage. Three programs of greedy algorithm, ascent algorithm and genetic algorithm based on MATLAB 7.0 are developed to solve the model. A computational experiment is adopted to prove the effectiveness and performance of model and heuristic algorithms. The results of simulation prove that the performance of gradual coverage model is better than traditional binary coverage model. Different weight coefficients of objective can affect the location results obviously. Twenty-five different scales data is used to compare the performance of heuristic algorithms, the performance of genetic algorithm is better than ascent algorithm, and greedy algorithm is worst. The ascent algorithm is suitable for the small and medium-sized location problems, and genetic algorithm is suitable for solving large-scale location problems.

Key words: emergency facility location, multi-objective, multiple gradual coverage, heuristic algorithms

摘要： 大规模突发事件下应急物资的需求量巨大以及对资源持续需求的特点，考虑设施选址的公平性、效率性及成本等因素，基于多级覆盖和覆盖衰减思想，提出一类应急设施多目标多重覆盖衰减选址模型。基于MATLAB7.0设计贪婪算法、上升算法、遗传算法程序对模型进行求解，以25组不同规模的算例验证了模型的性能和算法的有效性。数值模拟结果表明，该模型较之传统覆盖选址模型可以为需求点提供更高的覆盖满意度；当目标权系数取不同值时对选址结果产生较大影响；对三个算法性能进行比较，遗传算法最优，上升算法次之，贪婪算法最差，上升算法适于求解中小规模的选址问题，而遗传算法更适合于大规模选址问题的求解。

关键词: 应急设施选址, 多目标, 多级覆盖衰减, 启发式算法