Mathematical models and solution algorithms for the vehicle routing problem with environmental considerations

Raeesi, Ramin and Zografos, Konstantinos (2019) Mathematical models and solution algorithms for the vehicle routing problem with environmental considerations. PhD thesis, Lancaster University.

Text (2019raeesiphd)
2019raeesiphd.pdf - Published Version
Available under License Creative Commons Attribution-NoDerivs.
Download (4MB)

Abstract

Urban freight distribution is essential for the functioning of urban economies. However, it is contributing significantly to problems such as traffic congestion and environmental pollution. The main goal of this research is to contribute to greening urban freight distribution by developing new mathematical models and solution algorithms pertaining to the two major steams in Vehicle Routing Problems (VRPs) with environmental considerations: (i) VRPs with an explicit fuel consumption estimation component as a proxy for emissions, and (ii) VRPs with vehicles in the fleet that run on a cleaner alternative fuel such as electricity. In the first stream, this thesis develops and solves a new realistic multi-objective variant of the pollution-routing problem, referred to as the Steiner Pollution-Routing Problem (SPRP), that is studied directly on the original urban roadway network. The proposed variant is capable of incorporating the real operating conditions of urban freight distribution, and striking a balance between traditional business and environmental objectives, while integrating all factors that have a major impact on fuel consumption, including the time-varying congestion speed, vehicle load, vehicle’s physical and mechanical characteristics, and acceleration and deceleration rates. The thesis develops new combinatorial results that facilitate problem solution on the original roadway network and also introduces new mathematical models for synthesizing the expected second-by-second driving cycle of a vehicle over a given road-link at a given time of the day. New efficient multi-objective optimisation heuristics are also developed for addressing realistic instances of the SPRP. On the other hand, in the latter stream discussed above, to tackle the significantly impeding problem of range anxiety in the face of goods distribution using Electric Commercial Vehicles (ECVs), a paradigm shift in the routing of ECVs is proposed by introducing the Electric Vehicle Routing Problem with Synchronised Ambulant Battery Swapping/Recharging (EVRP-SABS). The proposed problem exploits new technological developments corresponding to the possibility of mobile battery swapping (or recharging) of ECVs using a Battery Swapping Van (BSV). In the EVRP-SABS, routing takes place in two levels for the ECVs that carry out delivery tasks, and for the BSVs that provide the running ECVs with fully charged batteries on their route. There is, therefore, a need to establish temporal and spatial synchronisations between the vehicles in the two levels and to do so new combinatorial results and a new solution algorithm is proposed.