A model for three staged reverse distribution systems and its decomposed solution

Reverse logistics is gradually gaining importance due to the purpose of sustainable and profitable use of resources and government legislations. For the purpose of using resources efficiently by recycling generic models are needed. In this thesis, a new reverse distribution model that combines forward and reverse distribution in a recovery network is submitted. The model that is presented is a three staged, capacitated network model including the costs of opening facilities, carrying stock and the costs of production, handling, transportation, collecting and disposal of products. The model includes production/recovery/dosposal facilities in its first stage, distribution/collection facilities in its second stage, sub distribution/collection facilities in its third stage, and customers in its fourth stage. It is the first cost minimizing generic recovery model that has three stages, capacities and carrying stock developed in the literature. The costs that are in the paths from production facility to customer and from customer to the recovery facility, belonging to all of the stages, are combined and simplified. Optimal solution techniques are insufficient for the derived problems as the stages and the variables of model are increasing. Heuristics can not produce optimal solution. It has been showed that the constraints of the model are convenient to decomposition method and an application is carried out optimally by Dantzig-Wolfe decomposition method. Decomposition algorithm and flow diagram are presented. It is demonstrated that forward flow and the facilities in that flow form a sub problem, backward flow and the facilities in that flow form another sub problem, and balance equation of the model combines two of the sub problems.