Modeling decentralized supply chain with centralized and decentralized approaches

摘要

In this paper, a multistage decentralized supply chain (SC) with different independent entities is modeled with the combination of centralized and decentralized decision-making processes. To do so, mathematical models are developed, based on the concepts of bilevel programming and the weighted goal programming approach. Here, a strategy of centralized SC control is proposed as an upper level (leader) optimization problem, whereas the individual profit maximization problems of each of the independent SC stages are considered as lower level (follower) problems under decentralized control. To implement the bilevel programming concept, an optimization model is developed as a leader problem to coordinate different independent entities of a decentralized SC and a few independent optimization models for different SC stages are considered for the follower problems. We have introduced a solution approach to solve the proposed multiperiod model using a goal programming approach, where a target/goal value is set for each lower level optimization problem to obtain feasible solutions of the proposed bilevel model. A numerical analysis is conducted to validate our model. Moreover, our proposed model is compared with a single-level approach, where total SC profit maximization is considered, instead of optimizing each individual entities' problems. The comparison shows that modeling a complete decentralized SC network with independent entities having individual objectives with our proposed model is more realistic and effective than single-level modeling.