Battery-time-space fragment-based formulation for the Electric Autonomous Dial-A-Ride Problem
Résumé
The Electric Autonomous Dial-A-Ride Problem (E-ADARP) focuses on optimizing the routes and schedules for a fleet of electric autonomous vehicles to transport a set of customers from their origins to destinations. E-ADARP integrates two primary features: (i). a combined objective that addresses both total travel cost and excess user ride cost; (ii). the use of electric autonomous vehicles that incorporate partial recharging, thereby linking the charging level to the time consumed for charging. This paper presents a battery-time-space fragment-based formulation (BTSFBF) to tackle the Electric Autonomous Dial-A-Ride Problem (E-ADARP). In this model, a fragment refers to a segment of the route, with each segment assigned a cost parameter that accounts for both the total travel cost and the excess user ride cost for customers within that segment. Additionally, BTSFBF utilizes a battery-time-space network where time and battery levels are discretized into a finite number of indices. In this network, actual time and battery consumption for each route fragment are rounded down to the nearest discrete interval. To address potential inaccuracies from the rounding, when solving BTSFBF, a callback mechanism is employed to eliminate infeasible paths, thereby ensuring feasible routing solutions. The computational experiments indicate that the BTSFBF method performs better than the traditional branch and price algorithm in terms of solution time and quality.