Combining affinity propagation with differential evolution for three-echelon logistics distribution optimization

In order to alleviate urban congestion, improve vehicle mobility and logistics distribution efficiency, the urban logistics distribution system is regarded as a three-echelon logistics distribution system. In this paper, a mathematical model of the 3-echelon logistics distribution problem (3E-LDP) considering time window constraint is established based on the directed graph, and a double-tier intelligent algo-rithm solution scheme is proposed, which combines the distance entropy-based Affinity Propagation clustering (DEBAP) algorithm and the crossover and selection-based differential evolution algorithm (CSBDE). First of all, in order to reduce the scale of logistics distribution and improve the utilization rate of logistics distribution facilities, the DEBAP algorithm is proposed in the upper tier to divide the logistics distribution region and optimize the distribution of logistics facilities, and the resulting scheme is passed to the vehicle routing optimization algorithm in the lower tier. Secondly, the vehicle routes at all levels are optimized based on the CSBDE algorithm at the lower tier, and the optimized route scheme is fed back to the DEBAP algorithm at the upper tier, so as to coordinate multi-echelon logistics distribution. Then, a search strategy based on the reachable distribution region and a facility allocation optimization strategy based on the weight of routing length are proposed to improve the efficiency of the algorithm. Based on the above algorithms, the optimization of the three-echelon logistics distribution system is completed in coordination. Finally, the performance of the proposed method is evaluated on the standard benchmark instances of the problem. The experimental results show that the three-echelon logistics model can improve the efficiency of logistics distribution, and the method has the best comprehensive performance, which is better than the most advanced 3E-LDP solution method. It has great potential in practical projects. (c) 2022 Elsevier B.V. All rights reserved.