A Multiobjective Evolutionary Algorithm for the Bus Stops Location Problem: Optimizing Demand Coverage, Travel Times, and Operational Costs

Abstract

Public transportation is a key component of urban mobility, providing accessible and efficient services for citizens. However, proper planning is essential to ensure both quality of service for users and cost-efficiency for system managers. In this context, this article addresses the Bus Stops Location Problem, which focuses on determining optimal locations for bus stops to balance user accessibility, service efficiency, and operational costs. A multiobjective evolutionary algorithm is devised to address the problem. The formulated problem model simultaneously optimizes three objectives: (i) maximization of demand coverage, (ii) minimization of travel times (approximated by reducing the number of stops), and (iii) minimization of operational costs (weighted by demand). A real case study is addressed using public transportation mobility data for Montevideo, Uruguay. The proposed algorithm is implemented as an integer multiobjective problem using the jMetal framework. Experimental results indicate that the proposed multiobjective evolutionary algorithm reliably produces high-quality Pareto fronts capturing meaningful trade-offs between coverage, travel time, and cost: solutions tend to prioritize high-demand central areas while offering diverse alternatives to policymakers. The evolutionary approach attained accurate trade-offs between problem objectives, with comparable or lower travel time and cost for similar coverage levels than the current situation in the Montevideo case study. Travel times were improved by more than 30% and operational costs by 5%, while maintaining demand coverage. Moreover, the evolutionary approach achieved these results in less than half an hour on average, highlighting the high computational efficiency of the metaheuristic solution.