基于自适应遗传算法的应急物资储备库选址及物资调配优化研究

Approach to optimizing the location &allocation of the emergency material reserve based on the adaptive genetic algorithm

应急物资储备库是应急救援的重要载体,储备库选址是否科学合理决定着整个应急公共服务体系的成效,而传统遗传算法在研究储备库选址与物资调配相结合的问题上存在局限性。针对这一问题,从需求点及应急设施服务质量视角构建基于覆盖满意度和经济性的应急设施选址与物资调配优化模型,并根据模型设计自适应遗传算法(AGA),结合储备库实际情况采取具有方向性的初始群体生成法提高搜索速度,设计自适应交叉与变异算子使AGA能够在进化速度与解的质量之间进行权衡,并获得全局最优解。结果表明,该算法能跳出局部最优从而获得最优的选址-分配方案,可为应急管理部门和决策机构的应急设施选址规划及应急物资分配提供理论依据和决策参考。

The purpose of this paper is to establish a multi-objective model for optimizing the location selection of the emergency material and facilities so as to provide for preliminary countermeasures and suggestions for the decision makers to follow in accident prevention. To achieve the purpose,first of all,the paper has to build up first of all an emergency facility location-allocation optimization model with the time-effective and cost-effective objectives to coordinate both with the allocation and schedule of such materials. And,secondly,it is necessary to determine the constraints to ensure the rationality of the solution. At the same time,the problem on how to solve the traditional genetic algorithm( GA) helps to optimize the solvation of the adaptive genetic algorithm( AGA) locally. And,therefore,in accordance with the demands of the model,it is necessary to design the chromosome structure to express the location-scheduling project. And,thirdly,to improve the quality of the initial population,a directional initial population generation method has to be adopted to solve the problem that the genetic algorithm is highly dependent on the initial population so as to improve the searching speed.And,fourthly,it is also to balance the evolution speed of the adaptive design of the crossover means and enable the AGA to solve the problem of the fixed crossover rate and the mutation rate parameter,for they are likely to affect seriously the quality of the solution through the mutation operator,so as to eventually verify the case sample and compare it with the genetic algorithm and justify the effectiveness of the method proposed. Thus,it can be seen that the results we have gained may suggest that,under the same setting of parameters,the traditional algorithm can not jump over after being trapped in the local optimum,and,then,higher population fitness can be obtained by the designated AGA in the initial population and continuously jump out of the local optimal mode, so as to obtain the optimal siting-allocation scheme at 70 iterations. And,therefore,the location of the storage warehouse and the allocation of the materials in accordance with the model we have put forward can be taken as the justified one to provide for a comprehensive solution to some extent.