Hybrid optimization of logistics distribution paths using neural networks and fuzzy logic

In the context of the booming logistics industry, where the scarcity of integrated applications of multiple intelligent technologies in logistics distribution path optimization is prominent, this paper innovatively presents a hybrid optimization strategy that combines neural networks and fuzzy logic. This strategy doesn’t merely integrate the two technologies; it fills the research gap in multi - intelligent - technology collaborative optimization of distribution routes in complex logistics environments. In the experimental phase, in addition to the commonly - used TSPLIB, CVRPLIB, OR - Library, GEO - TSP, and VLSITSP datasets for performance testing, a vast amount of real - world logistics - related data, including historical distribution path details, traffic flow data, and customer demand information, is collected for neural network training. Rigorous pre - processing is carried out on the training data, such as cleaning to remove outliers and normalization to unify data scales. The proposed model is compared with classic algorithms like the greedy algorithm, genetic algorithm, particle swarm optimization algorithm (PSO), and Dijkstra algorithm. The experimental results demonstrate that the proposed model outperforms others in multiple crucial aspects. It attains the shortest delivery time, the lowest delivery cost, the highest customer satisfaction, the shortest path length, and the lowest robustness index value across all datasets. These outcomes indicate that the proposed model not only boosts distribution efficiency but also strengthens system stability and reliability, holding substantial practical application value in the logistics field.