Optimal design of a multi-period supply chain network with unreliable elements: A two-stage stochastic programming approach

The management and design of supply chain networks in various dimensions are so critical today that managers' decisions significantly impact the configuration and flow of material in the network. Above all, supply chain management intends to re- duce costs. The inability to accurately predict certain features, such as demand, can complicate the cost estimation process. To that end, an essential parameter is the reliability of supply chain networks. Considering the reliability of the supply chain network brings the model closer to reality, and the wellness or failure of its elements under different scenarios increases the enthusiasm to face unpredictable events in managers and helps network performance. Furthermore, appropriate management and design of the supply chain network can increase customer satisfaction and reduce costs in the long term. In this research, a four-tier supply chain network was designed to reduce the costs through a two-stage stochastic programming attitude. The combined metaheu- ristic method (genetic and simulated annealing algorithms) was used to solve the model. By treating the reliability of entities and routes and its effect on reducing cost as an essential criterion in the mentioned problem, it was showed that a reliable system has lower costs than an unreliable system.