基于多联式遗传算法的运输网络成本研究  

Research on cost of transport network based on multi-type genetic algorithm

作　　者：杨璐[1] 唐勇军[1] 

机构地区：[1]河海大学商学院,江苏南京211100

出　　处：《电子设计工程》2017年第9期153-156,161,共5页Electronic Design Engineering

基　　金：国家自然科学基金资助项目(71102072);国家社会科学基金资助项目(15BGL054)

摘　　要：针对目前多联式运输网络成本问题,本研究通过使用虚拟网络建模综合考虑了公路、铁路和内河航运的网络子节点连接,并构建了4种运输链接模式来计算广义成本。利用全有或全无的算法(AON)用于路径流量的分配,在Michalewicz开发的编程逻辑基础上使用GISDK中的二次开发环境创建了自脚本,提出了针对运输网络成本的多联式遗传算法。该算法在上层求解运输网络模型的最优解。通过实验仿真,结果显示:在处理运输网络成本过程中,多联式遗传算法误差率有效的控制在8%以内;相比传统遗传算法,多联式遗传算法经过AON路径优化后的货物流量分配加速了算法的收敛效率,可以产生一个相对良好的运输路径网络,且性能稳定。Aiming at the multi-modal transportation network costs, this study by using a virtual network modeling considering the road, rail and inland shipping child node network connection, transport links and builds four models to calculate the generalized cost. Use all or nothing algorithm （AON） for dispensing flow path using GISDK the secondary development environment programming logic developed on the basis of Michalewicz created from a script, a multi-connected transportation network costs for genetic algorithm. The algorithm for solving the optimal solution in the upper transport network model. The simulation results show that： Processing costs during transportation network, multi-type genetic algorithm error rate effectively controlled within 8%; compared to traditional genetic algorithm, multi- type genetic algorithm convergence efficiency and stable performance.

关 键 词：遗传算法 网络子节点 运输网络 运输成本 

分 类 号：TN02[电子电信—物理电子学]