Solving High Volume Capacitated Vehicle Routing Problem with Time Windows using Recursive-DBSCAN clustering algorithm (1812.02300v2)

Abstract: This paper introduces a new approach to improve the performance of the Capacitated Vehicle Routing Problem with Time Windows (CVRPTW) solvers for a high number of nodes. It proposes to cluster nodes together using Recursive-DBSCAN - an algorithm that recursively applies DBSCAN until clusters below the preset maximum number of nodes are obtained. That approach leads to 61% decrease in runtimes of the CVRPTW solver as benchmarked against Google Optimization Tools, while the difference of total distance and number of vehicles used by found solutions is below 7%. The improvement of runtimes with the Recursive-DBSCAN method is because of splitting the node-set into constituent clusters, which limits the number of solutions checked by the solver, consequently reducing the runtime. The proposed method consumes less memory and is able to find solutions for problems up to 5000 nodes, while the baseline Google Optimisation Tools solves problems up to 2000 nodes.