Vehicle routing in a forestry commissioning operation using ant colony optimisation

Abstract

This paper formulates a vehicle routing problem where constraints have been produced from a real world forestry commissioning dataset. In the problem, vehicles are required to fully load wood from forests and then deliver the wood to sawmills. The constraints include time windows and loading bay constraints at forests and sawmills. The loading bay constraints are examples of inter-route constraints that have not been studied in the literature as much as intra-route constraints. Inter-route constraints are constraints that cause dependencies between vehicles such that more than one vehicle is required to perform a task. Some locations have a lot of consignments at similar times, causing vehicles to queue for loading bays. The aim is to produce an optimal routing of consignments for vehicles such that the total time is minimised and there is as little queuing at forests and sawmills as possible. In this paper, the problem has been formulated into a vehicle routing problem with time windows and extra inter-route constraints. An ant colony optimisation heuristic is applied to the datasets and yields feasible solutions that appropriately use the loading bays. A number of methods of handling the inter-route constraints are also tested. It is shown that incorporating the delay times at loading bays into the ant’s visibility produces solutions with the best objective values.