一种基于遗传算法的多模式多标准路径规划方法

单一路径评价标准难以满足日益多样化的出行路径规划需求。多标准路径规划成为公众出行服务的研究热点。然而，多标准路径规划本质上是具有NP特性的多标准决策问题，且涉及多种交通出行模式。多个不同标准的权重设置将直接影响路径规划结果。因此，如何科学合理地设置不同标准的权重成为多标准路径规划中的技术瓶颈。本文提出了一种适应多模式交通网络环境的多标准路径规划方法，借鉴遗传算法在求解多标准优化问题中的优势，将其扩展到多模式多标准路径规划中。该方法避免了不同出行标准权重设置中的主观性和不确定性，能够实现更为灵活的交通出行模式自动化组合，为出行者提供满足个性化需求的、多标准的出行路径规划服务。

A multi-modal multi-criteria route planning method based on genetic algorithm

Single criterion routing can not meet the diverse needs of individuals for travel path selection. How to provide multi-criteria routing service has been a hot topic for advanced travel information systems. However, the multi-criteria routing is basically a complex NP problem, and involves different transportation modes. Arbitrary weight assignment for various criterion will remarkably affect the routing results. This paper proposes a multi-criteria route planning method for multi-modal transportation system. It takes the advantage of genetic algorithm for solving optimization problems and extends it to multi-modal routing environment. Various length of chromosome with mode tags is used to encode individuals. Both intra- and inter-mode evaluation operators are defined to guarantee the diversity evaluation. Pareto ranking method with a p-dimensional vector representing multiple criteria is used for fitness calculation. The presented method avoids subjective weight setting procedure, and can obtain various modes combination results for route planning to meet personalized requirements.