A hierarchical multiobjective problem for green logistics

Abstract

Throughout all the extension of this thesis a situation where a distribution company and a manufacturer company work within a supply chain in a hierarchically manner. Each one of them has to optimize their own operations to minimize costs in the case of the manufacturer company and maximize profit as well as minimize CO2 emissions for the distribution company. A green logistic approach is also desired in this research, which is one of the reasons that a minimization of greenhouse gas emissions is considered. The resulting formulation of the problem is a multiobjective bilevel model, therefore a GRASP metaheuristic within a weighted sum method was developed to tackle this problem. In addition, a second simpler methodology which consists to reformulate the problem only as a selection of the customers that will be visited was used. In the latter methodology, the specific modeling known as the Multiple Choice Model was implemented to remove the bilevel aspect of the situation resulting in an easier model. Afterwards, this model was solved by a proposed Labeling Algorithm which will be compared with the traditional -constraint method for multiobjective problems. Twelve test instances were generated in order to assess both methodologies. The experimental results as well as the approximated Pareto front are shown. Finally, this thesis work presents the importance of using these kind of modeling techniques in a supply chain.