Route Network Design of Community Shuttle for Metro Stations Through Genetic Algorithm Optimization

Abstract

This paper investigates an issue for optimizing community shuttle network linked with metro stations. Considering a set of trip demands originating from several bus stops within a large community area, a mixed integer optimization model is formulated to obtain the optimized shuttle route network and the frequency of each route simultaneously, to minimize the total transit system cost, including user and supplier costs. To solve the problem, a solution framework consisting of three primary components is proposed: an initial route network scheme is developed to ensure that any generated network is feasible; a network analysis procedure is performed that assigns the demand of each stop to a set of paths on the generated network according to a certain proportion; a heuristic algorithm is proposed to optimize the service frequency of each route; a genetic algorithm consisting of a set of specifically designed operators is proposed to guide the solution population evolving process. A real-life study is tested successfully using the proposed solution framework. The related sensitivity analysis and impacts of some important variables on the solution are performed.