Disruption management for the electric vehicle routing problem in a geographically flexible network

Abstract

The adoption of electric vehicles (EVs) in populated cities is increasing in response to reducing the environmental impact of transportation; however, it also brings new problems to be solved that require innovative solutions such as the optimal use of charged vehicles, the location of charging stations, and managing transportation disruptions in a dynamic environment. This study addresses the problem of recovering a pre-established schedule of an EV when an unexpected disruption occurs. The innovative idea for this problem is reconfiguring the road network by skipping one or more customers while locating alternative points for the temporary storage of consignments initially scheduled to be picked up from (or delivered to) skipped destinations. It is allowed to assign lockers to a neighborhood entailing a set of nodes. This paper designs an integrated form of EV routing problem that simultaneously determines the optimal velocity in each assigned route and the battery recharging policy under a partial charging scheme. The paper also proposes an efficient algorithm based on the Crowd-Learning Particle Swarm Optimization (CLPSO) to solve the large-scale problem. The proposed algorithm surpassed a set of widely used algorithms in the literature using a numerical case study. Computational experiments, based on data from a freight company, demonstrate the effectiveness of the model and CLPSO algorithm. The test results confirm that the developed approach can be a useful reference in practice to provide a robust operation of EVs in metropolitans and populated cities. The results also show that applying all of the proposed recovery actions can significantly reduce the cost of the disruption.