Abstract: To facilitate selection of school bus stations, allocating student groups to stations and developing vehicle routings, the mathematical programming model was established to minimize the cost of the vehicle travel time, the students journey time, and the students walking time to the station by considering the key constraint conditions such as the longest riding time in the school bus and the longest walking time of students to the station. Then the improved genetic algorithm was put forward to for above purpose and by this algorithm, the excellent individual of initial population was produced by heuristic method, and the operators with heuristic knowledge were designed to improve the excellent searching efficiency. The example analysis results show that the proposed method is feasible with significant results, which can effectively provide optimal route for school bus of large number.

Key words: traffic and transportation engineering, school bus routing, optimization, mathematical programming model, improved genetic algorithm

摘要： 以校车站点选择、学生群归属站点的划分以及车辆路径安排为研究对象，重点考虑了学生在车上的最大乘车时间、学生步行到候车站点的最大步行时间等约束条件，以车辆行程时间成本、学生乘车时间成本以及学生步行时间成本最小为目标建立数学规划模型。给出了解决这类问题的改进遗传算法，该算法通过启发式产生初始种群的优良个体，并针对模型特点设计带启发知识的遗传算子，提高寻优效率。实例分析表明，该方法可行，并且有比较显著的效果，能够有效地解决大规模的校车路径优化问题。

关键词: 交通运输工程, 校车路径, 优化, 数学规划模型, 改进遗传算法