Using Deep Learning Artificial Intelligence to Solve the Capacitated Vehicle Routing Problem

Abstract

The vehicle routing problem is a complex optimization challenge that belong to the class of NP-Hard problems. This implies that it is challenging to find an optimal solution within Polynomial time or that the solution process may take a significant amount of time. The main objective of this problem is to minimize the total distance traveled by vehicles. In this research, to propose the use of deep learning artificial intelligence, specifically neural networks, to determine the most appropriate heuristic algorithm for resolving the vehicle routing problem which include with six algorithm, including 2-Optimize, 2-Approximate, Nearest_2Opt, Improve nearest, Rep improve nearest and OR tools. The clustering of vehicle routing is then performed using the Sweep algorithm. The sequence of the vehicle’s path is determined such that larger vehicle prioritize routes first, followed by smaller vehicles, as the use larger vehicles results in lower mileage than smaller vehicles. The result of our study demonstrate that when comparing the optimal distance from the standard dataset to the solved distance by artificial intelligence, the mean distance from artificial intelligence is greater than the optimal distance by no more than 17% at the 95% confidence level.