A new approach based on hybrid ant colony optimization-artificial bee colony algorithm for multi-objective electric vehicle routing problems

Abstract

In real-life, routing plans try to optimize multiple objectives without relying on just one. So, this paper introduces three multi-objective electric vehicle routing problems (MOEVRP) that consider different charging strategies and electric vehicle (EV) charger types while optimizing five conflicting objectives: a total mini-mization cost of recharging, the number of vehicles required, a total travel distance, load-dependent energy consumption, and the total number of charging stations required. We develop a new hierarchical approach consisting of two phases: a Hybrid Ant Colony Optimization (HACO) and an Artificial Bee Colony Algorithm (ABCA). In the first phase, an initial solution is obtained using a HACO that integrates local search algorithms and simulated annealing (SA) to reduce the solution time. Then in the second phase, the problem is solved using an ABCA considering the initial solution obtained from the first phase. Using the proposed HACO-ABCA as the search engine, two posteriors' methods, namely the weighted-sum method (WSM) and the conic method (CM), are applied to scalarize the five objectives. The effectiveness of the proposed hierarchical approach examined on well-known test-based instances and obtained the best new results in most instances. Additionally, the proposed solution is applied to a real-life case study. The results show that multi-objective traditional methods give more effective results than multi-objective evolutionary algorithms, regardless of the MOEVRP problem type. We can also conclude that the partial recharge and multiple recharge technology options can significantly improve the route decisions of logistic companies.