Nested Rollout Policy Adaptation for Optimizing Vehicle Selection in Complex VRPs

Abstract

The goal of Vehicle Routing Problems (VRP) and their variations is to transport a set of orders with the minimum number of vehicles at least cost. Most approaches are designed to solve specific problem variations independently whereas in real world applications, those constraints are handled concurrently. This paper describes a novel approach to solve variants of Open VRP, Multi Depot VRP, Capacitated VRP as well as Pickup and Delivery with Time Windows applying Monte-Carlo Tree Search (MCTS) and in particular Nested Rollout Policy Adaptation. For evaluation, real data from the industry was obtained and tested on the developed approach.