Green vehicle routing optimization based on dynamic constraint selection co-evolutionary algorithm

Aiming at the problems of single solution objective in the existing green vehicle routing optimization process and real-time speed change during vehicle travel, a multi-objective green vehicle routing problem with time window constraints in time-varying conditions(MOGVRPTW-TV) is established, then a co-evolutionary framework-based constrained multi-objective evolutionary algorithm for the solving of the model is proposed. First, to better match the actual logistics distribution, this model considers the impact of time-varying speed and capacitated variation on carbon emissions based on capacitated constraints and time window constraints. Second, a constrained multi-objective evolutionary algorithm based on the co-evolutionary framework was proposed for model solving. The algorithm treats the complete problem model as a complex task and introduces a shift crowding distance calculation that considers both individual distribution and convergence information when solving this complex task, effectively balancing the convergence and diversity of solutions. Then, a dynamic constraint selection strategy is designed for the implementation of the simple task, which only considers the effects of some constraints on the populations, and the two populations exchange information through offspring populations to achieve overall optimization. Simulations are performed on different instances and the results show that the proposed algorithm is effective in avoiding traffic congestion periods, decreasing overall distribution costs, and reducing fuel consumption and carbon emissions of vehicles.