Optimization of Public Transport Networks by Considering Alternative Positions for Network Stations

Abstract

Netheradays planning of metropolitan areas considers to improve quality of life of their inhabitants and urbanmobility is one of the main concerns. Studies point out that investments in public transportation and other modes are aimedat the overall improvement of mobility. However, there is a gap in proper tools for optimizing public transport networks.In fact, network optimization is an NP-Hard problem and there are usually many conflicting objectives that need to beoptimized simultaneously. This paper proposes the use of manyobjective evolutionary algorithms to address the problem ofpublic transport networks optimization, focusing on metropolitan bus lines. The proposal consists in optimizing the position of bus stops and consequently obtaining new routes that pass through these stops in order to minimize the average travel time, the time spent between origin / destination and the variance of distance between the stops. To evaluate our proposal, a simulator was used to simulate the behavior of different passenger profiles in an urban area and the results were compared between the lines obtained by the optimization process and existing bus lines in the city of Sao Paulo. According to our results, optimized bus routes have mean travel time 22% less than the existing route and the time spent between origin/destination has decreased up to 18%.