AbstractAlthough urban freight transportation plays an important role in economic growth, it also constitutes one of the largest sectors generating more externalities which include social and environmental impacts such as pollution, depletion of natural resources and accidents. Also, this sector deals with a high level of stochasticity in customer dependent and time related information. Given its key role in development, smart approaches for designing sustainable and robust routes are needed. This thesis focuses on the economic, environmental and social sustainability impacts of urban freight transportation. In order to investigate and address these problems, some novel optimisation models as well as efficient algorithms and methods to handle this complex rich problem are developed and implemented.
In this thesis, some original contributions are proposed. First, an extension of the deterministic sustainable vehicle routing problem with economic and environmental objectives and three algorithmic approaches to address the problem are proposed. This ‘realistic’ problem arises in practice especially when the social impacts of urban freight transportation are ignored. Social impacts and costs of logistics and transportation activities are often borne by the citizen, however, freight transportation companies should bear some of these costs. Therefore, a weighted sum optimisation model to also include a social objective is developed. An advantage of using a multi-objective model is to investigate robustness of solution by means of sensitivity analysis.
Secondly, when some information is not known in advance, deterministic models fail to take into account these uncertainties. For this reason, two novel recourse models are developed. The first model addresses travelling time uncertainties and tries to minimise the routing costs which includes the driver costs, fuel costs and penalty cost of driver overtime. The second model is a weighted sum recourse model that takes into account uncertain travelling times and demands and tries to minimise economic, environmental and social costs of freight distribution. Finally, a hybridisation of heuristics with Monte-Carlo Simulation to handle uncertainties is proposed.
Lastly, an extensive set of computational experiments are carried out to investigate the interrelationship between the sustainability dimensions, the effect of stochasticity on the sustainability indicators and the trade-off(s) among them.
To the best of our knowledge, this is the first collection of work that proposes and implements the described optimisation models and solution methods to solve the logistics and transportation distribution problem.
|Date of Award||Oct 2018|
|Supervisor||Djamila Ouelhadj (Supervisor) & Dylan Jones (Supervisor)|