面向防汛物资动态变化的运输车辆调度优化算法研究

为解决防汛救灾过程中受灾需求变化下的防汛物资调度问题,提出一种面向防汛物资动态变化的运输车辆调度优化算法(SOA_TV).在SOA_TV算法中,考虑车辆限载、调度车辆数量、移动距离等约束,建立防汛物资运力调度优化模型.依据已知受灾信息和仓储信息,确定受灾最小所需车辆数,获得待运输物资集合,并按照最近邻原则初始化车辆集合.引入车辆移动距离阈值,通过边权计算构建二分图,并进行矩阵转换,获得一个低维度的矩阵.最后,考虑需求不变化和动态变化两种情况下的物资分配,根据仓库点之间的运输距离和车辆负载情况更新边权值,多次执行KM算法直到获得目标模型的近似最优解.实验结果表明:在多种实验场景中,SOA_TV都能寻找到一个较优解.相比于GA和ABC,SOA_TV虽然略微降低了车辆移动总距离,但其运算时间获得大幅度削减,可在极短的时间内计算获得较优的车辆分配方案.相较于Hungarian,SOA_TV可降低运行时间和车辆移动总距离.

Scheduling optimization algorithm of transportation vehicles for dynamic changes of flood control materials

In order to solve the scheduling problem of flood control materials with the change of disaster demands in the process of flood control and disaster relief, this paper proposed a scheduling optimization algorithm of transportation vehicles(SOA_TV) for dynamic change of flood control materials. Considering the constraints of vehicle loads limit, the number of dispatched vehicles and the moving distance, SOA_TV established the capacity scheduling optimization model of flood control materials. According to the known disaster information and storage information, it determined the minimum number of vehicles required for disaster, and obtained the material set to be transported. Then it initialized the vehicle collection according to the nearest neighbor principle and constructed a bipartite graph through vehicle moving distance threshold and edge weight calculation. SOA_TV performed matrix conversion to obtain a low-dimensional matrix. Finally, considering the material distribution under the conditions of unchanged demand and dynamic changed demand, SOA_TV updated the edge weights according to the transportation distance between the warehouse points and the vehicle load, and executed KM algorithm many times until obtained the approximate optimal solution of the target model. The experimental results show that SOA_TV can find better solution in a variety of experimental scenarios. Compared with the GA and ABC, SOA_TV slightly reduces the total distance of vehicle movement, but greatly reduces its calculation time, and calculates a better vehicle allocation plane in a very short time. Compared with Hungarian, SOA_TV reduces the calculation time and the total distance of vehicle movement.