基于GA算法的轻轨车辆任务分节及配对优化研究

A GA-Based Algorithm for Light Railway Vehicle's Task Sectioning and Mating Optimization

针对轻轨车辆在指定两站进行任务分节具有多样性,论文采用人工智能直观启发确定了任务分节及配对的基本模型.建立了任务配对的目标函数和约束函数,保证不同车之间的2节任务配对后,换乘时上下车在同一站,时间在0.5至2小时之间,且使全部配对任务中换乘休息的总时间最短.设计了有效的“自交叉”GA算法,优化出各车的任务分节及其配对组合方法.算法对于日分多节任务、多换乘位置点的任务分节配对的排班问题具有一般适应性.此算法在天津轻轨运用管理信息系统中得到了良好的应用验证.

This paper analyses the characteristics of the light railway vehicle＇s running task, which is multiplicate when divided into some sections, and pattern shooting when paired two sections. We use the AI to determine the basic model of task＇s sectioning and mating, and build the object function and the constraint function. It assures that the driver gets off and gets on the train at the only one of two appointed stations, and the off-on-train＇s time is between 0.5 to 2 hours, and the total off-ontrain＇s time of the all drivers is minimum. We design a specific self-cross genetic algorithm to solve this problem, and get the optimum design of the task sectioning and mating. This algorithm can be used to the general problem that it should be divided into multi-section and off-on-train at the same station in the specific multi-station. We get a good result when this algorithm has been used in the MIS of the Tian-jin littoral express light railway.