基于免疫粒子群算法的多供应主体应急物资供应模型求解

Solution of Emergency Supply Model under Multi-supply Subjects Using an Immune Algorithm-based Particle Swarm Optimization

为了弥补政府应急物资库供应能力不足,提升救灾物资供应种类、时效和能力,提出由政府认证的协议供应商参与的多供应主体模式,并以供应偏差影响度最小和供应耗时最少为目标,构建了多目标多周期混合整数应急物资供应模型,以解决应急物资多种类动态供应问题。结合免疫信息运作机制和粒子群算法的免疫粒子群算法,以青海玉树地震物资供应数据为例,求出物资供应方案。分析了方案与实际供应数据对比差异,以及模型的有效性。结果表明:多供应主体参与的物资供应模式较传统物资库供应在救援及时性和需求满足度上有明显优势;免疫粒子群算法对多供应主体模型的求解时间较CPLEX求解器有大幅缩短,提升了模型在实际场景中应用的可操作性。

In order to make up the shortfall of limited capacity of government＇s emergency material warehouses for the sudden blowout demand in a major nature disaster,and improve the type,the effectiveness and the capacity of relief materials supplying,a supply mode of multi-supply subjects participated by the agreement suppliers with government certification is proposed,and a multi-objective multi-cycle mixed integer model of emergency supplies is developed,with the aims of the minimum supply deviation influence and minimal supplying time to solve the multi-type emergency material dynamic supply problem.By using an immune algorithm-based particle swarm optimization（IA-PSO） which coupling the immune information processing mechanism with the particle swarm optimization,and taking the emergency material supply data for Yushu earthquake for example,the solution of emergency material supply is worked out.The difference between the supply plan and the real case as well as the effectiveness of the model are analyzed.The result shows that（1） the material supplying mode participated by multi-supply subjects has obvious advantages in the rescue timelines and the demand fulfillment compared with the traditional material supply mode;（2） the IA-PSO approach,compared with CPLEX sovler,can significantly shorten the resolving time to the supply model of multi-supply subjects and improve the operability of the model in the actual scene.