No abstract

In this research, we look for the possible reduction in vehicle kilometers that can be obtained when a cooperation of delivery companies has the ability to suggest network improvements to the local government. Different network changes are considered in our research: re-opening existing roads for the vehicles of the cooperation, widening roads in the network or converting existing roads into a one-way road with a higher speed. To find the best set of improvements given a fixed budget in a realistic road network, an Adaptive Large Neighborhood Search (ALNS) is proposed. Both the destroy and repair methods in this ALNS are unique for this problem. In order to get an indication of the possible reduction in vehicle kilometers and to test the performance of the heuristic, experiments on a set of 16 benchmark instances are executed. These benchmark instances are generated from a realistic city road network. Based on these experimental results, we can conclude that a set of 1 up to 4 network improvements can lead to a reduction in vehicle kilometers of on average around 2.4% over these 16 benchmark instances, while implementing a set of 2 up to 9 improvements can lead to a reduction of on average around 3.3% over this set of benchmark instances.