چكيده به لاتين
A robust bi-level single-product multi-period network design model is proposed for a competitive green closed-loop supply chain considering inventory and pricing decisions under uncertainty and disruption risks. The bi-level programming approach is used through this model to demonstrate the competition among two competing supply chains referred as leader and follower in the competitive market. The concept of stacklberg game has been implemented to model the competition in a way that in the upper level, leader as the main decision maker aims to maximize his profit, and in the lower level follower wants to cover his own goals. Then, after modeling the competition and applying pricing decisions by defining a price-dependent demand function with fuzzy price coefficients, disruption risks are analyzed through the model.
The proposed model considers the demand uncertainty and supplier disruption risks, simultaneously and is capable of dealing with these uncertainties by implementing resilience strategies including, inventory decisions, and having contract with reliable suppliers. Moreover, to consider the environmental issues, controlling the amount of CO2 emissions by the machineries in production centers should be controlled and adding a reverse flow to the supply chain network should be added. The possibilistic programming approach is used to deal with the associated uncertainties in the demand function. Then, the Karush-Kuhan-Tucker (K-K-T) optimality conditions are applied for converting the proposed bi-level model into a single-level equivalent form.
As the model aims to maximize the total profit and minimize the amount of CO2 emissions, simultaneously, the integrated model follows is bi-objective. Since a bi-objective model has been used to depict the leader objectives more clearly, the e-constraint method is implemented to make a single objective integrated model. Finally, the important managerial insights are obtained from/through an industrial case problem. Verification of the model and proposed solution method have been done in small scale and by changing some parameters the model's behavior have been examined, then the model accuracy was proved. The model presented in this thesis was considered in a real situation for solving a problem of Sepana Palaye Pars company. Therefore, the suggested model is implemented on this company by using real data and some analysis is proposed for improving the current situation of this industrial company.
