MULTI-OBJECTIVE EVOLUTIONARY ALGORITHM FOR THE TRANSIT NETWORK DESIGN PROBLEM

The transit network design problem (TNDP) aims to find a set of routes and corresponding frequencies for an urban public transportation system. We model the TNDP as a multi-objective combinatorial optimization problem whose resolution involves finding a Pareto front that represents different trade-off levels between opposite objectives, the travel and waiting times and the required fleet of buses. We propose a Multi-Objective Evolutionary Algorithm (MOEA-TNDP) which has three algorithmic components, 1) an algorithm that generates a pool of candidate routes, 2) an assignment model that evaluates a given solution and checks the fulfillment of constraints, 3) an evolutionary process based on the Non-Dominated Sorting Genetic Algorithm version II that produces an approximated Pareto front. MOEA-TNDP is tested with the case of Rivera (a medium/small size city of 65000 inhabitants), taking its real solution and the solutions generated by the GRASP-TNDP algorithm as a reference for comparison