Multi-swarm PSO based on Elite Opposite Learning on Electric Vehicle Routing Problem with Time Window

Electric vehicle routing problem with time windows(EVRPTW) is a new problem in the field of transportation, and it is difficult to solve with the traditional exact solution method. Heuristic algorithms are usually used to solve EVRPTW and obtain an approximate optimal solution. We use mixed-integer linear programming to model EVRPTW, taking into account various factors encountered in practice. As a heuristic solution algorithm, multi-swarm particle swarm optimization algorithm(M-PSO-EL) based on elite opposite learning has faster convergence speed and searchability for high-dimensional complex problems to solve EVRPTW. Design the charging strategy of electric vehicles based on EVRPTW. Genetic Algorithm(GA), Discrete Quantum Behavioral Particle Swarm Optimization (DQPSO), and M-PSO-EL are tested on Solomon's benchmark which contains different types of customer point distribution. The experimental results show that M-PSO-EL can effectively solve EVRPTW with 100 customer points.