A Two-Phase Optimization Model for Autonomous Electric Customized Bus Service Design

Motivated by the requirements of highly effective customized bus (CB) service and by the rapid growth of autonomous electric vehicles (AEVs), this paper studies a new optimization model for the autonomous electric customized bus (AECB) service, aiming at minimizing operating costs and improving vehicles' efficient use. The proposed model contains two phases: (i) optimization of the vehicle routing, charging operation and passenger-to-vehicle assignment for the fixed travel demands, and (ii) re-optimization of the service according to real-time dynamic travel requests. A solution approach is developed to address the proposed model based on adaptive large neighborhood search (ALNS). The extensive empirical analysis, conducted by considering real-world data on a large-scale instance, demonstrates the efficiency of the proposed approach and the quality of the generated solutions.