Knowledge Graph-Based Dynamic Differential Evolution Algorithm for E-Waste Recycling Vehicle Routing Problem

The optimization of E-waste recycling vehicle routing is challenging due to the potential for the emerge of new orders or the change of existing orders during the recycling process. To address the issue of recycling efficiency, a knowledge graph-based dynamic differential evolution algorithm (DE-KG) is proposed. Knowledge graphs are constructed for different delivery scenarios to facilitate data transmission and storage. An improved dynamic time warping is employed to calculate similarities between delivery scenarios for clustering analysis. Optimal solutions of similar delivery scenarios are traced and incorporated into the initial population of the algorithm, enabling rapid adaptation to dynamic changes. The relationships between constraints, objective functions, and individual solutions are stored in the knowledge graph, allowing for adaptive parameter adjustments to enhance the efficiency and robustness of route optimization. Experimental results demonstrate that the proposed method is effectively adapted to dynamic environments, significantly improving scheduling performance and recycling efficiency.