Optimization of a demand responsive transport service using multi-objective evolutionary algorithms

Abstract

This paper addresses the problem of optimizing a Demand Responsive Transport (DRT) service. A DRT is a flexible transportation service that provides on-demand transport for users who formulate requests specifying desired locations and times of pick-up and delivery. The vehicle routing and scheduling procedures are performed based on a set of requests. This problem is modeled as a multi-objective Dial-a-Ride problem (DARP), in which a set of objectives related to costs and user inconvenience is optimized while respecting a set of constraints imposed by the passengers and vehicles, as time windows and capacity. The resulting model is solved by means of three Multi-objective Evolutionary Algorithms (MOEA) associated with feasibility-preserving operators. Computational experiments were performed on benchmark instances and the results were analyzed by means of performance quality indicators widely used for multi-objective algorithms comparison. The proposed approaches demonstrate efficient and higher performance when optimizing this DRT service compared to another algorithm from the literature.