RESEARCH ON COMBINING GENETIC ALGORITHM (GA) AND DECISION TREE TO SOLVE VEHICLE ROUTING PROBLEM (VRP)

Abstract

This paper proposes a method combining Genetic Algorithm (GA) and Decision Tree (DT) to solve the Vehicle Routing Problem (VRP). The method uses DT to predict the number of vehicles needed, helping to improve the efficiency of population initialization in GA. Experimental results on the Solomon benchmark dataset (including Rxx, Cxx and RCxx) show that the combined GA + DT method provides high-quality solutions, small errors and reasonable computation time.