A Dual-Population-Based Co-Evolutionary Algorithm for Capacitated Electric Vehicle Routing Problems

• Published in: IEEE transactions on transportation electrification Vol. 10; no. 2; pp. 2663 - 2676
• Main Authors: Wang, Chao, Qin, Fang, Xiang, Xiaoshu, Jiang, Hao, Zhang, Xingyi
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 01.06.2024; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Ant colony optimization; Ant colony optimization (ACO); charging decision problem; Charging stations; co-evolution; Electric power demand; electric vehicle (EV) routing problem; Electric vehicle charging; Electric vehicles; Evolutionary algorithms; genetic algorithm (GA); Genetic algorithms; Heuristic algorithms; Optimization; Polynomials; Populations; Routing; Search problems; Social factors; Statistics; Transportation; Vehicle routing
• ISSN: 2332-7782; 2577-4212; 2332-7782
• DOI: 10.1109/TTE.2023.3294588

The capacitated electric vehicle (EV) routing problem is a challenging nondeterministic polynomial hard (NP-hard) problem consisting of two interdependent subproblems, the routing optimization problem and the charging decision problem. The routing optimization for EVs with a limited driving range is dependent on the available charging stations, while the charging decision is based on the charging demand that is estimated on the fixed route in return. Considering this coupling relationship, this article proposes a dual-population-based co-evolutionary algorithm (DPCA) that uses two evolution populations to collaboratively optimize these two subproblems. In the routing population, the charging station is regarded as a kind of customer with no demand, and an improved ant colony optimization (ACO) algorithm is designed to generate routes that involve the position information of charging stations. In the charging population, a binary genetic algorithm (GA) is used to generate a population of charging schemes whose qualities are evaluated based on the best ant obtained from the routing population, and then the resultant solution by inserting the best charging scheme is used to update the pheromone for the routing generation. Through the information interaction during the evolution, these two populations collaboratively search for the optimal solution to the problem. Experimental results demonstrate that the proposed algorithm can be able to avoid falling into the local optimum and have a reduction of about 4% in route distance averaged over two test suites. In addition, it also has a high computational efficiency, which is faster than the advanced ACO method by about two times.