Robust Optimization Of Supply Chain Management System Under Uncertain Conditions

Improving the performance and optimization all the processes of the supply chain network are the objection of our research work. Because of the complex net configuration of the supply chain, uncertainties is natural in the supply chain. In this paper, we analyze the resource of the uncertainties, review the optimization mothods at first. The second, we focused on several key problems in supply chain management in uncetainty conditions, studied these three problems:supply chain network construction, procurement production and distribution synchronizing, production inventory and transportation coordinating. We also constructed robust optimization models for the three problems respectively, design intelligent algorithms for them. Finally, through a prototypic system, we show the effectiveness of the models and their algorithms.The main originality ideas are those:(1) Not only the operation cost of supply chain but also the robustness of the system are optimized. So the sensitivity to the perturbation of the uncertain demand and the cost of supply chain both are lowered and the problem of difficult to determine the cost of stock out is avoid.(2) Taking into account the uncertainties of cost and demand, we prosented a robust optimization model of supply chain design. In the model, not only facility decision problem but also supplier selection problem are considered. We also present a Intelligent algorithm for the model. The numarical experiment shows that the robust optimization model not only ruduce the risk of the uncertain market and cost effectively but also avoid the error bucause of the difficulty of determining the stockout cost when modeling. Analyzed the influence to the system performance of the objective function risk coefficientλ. Study shows that robust solution is possible, but at some cost.(3) Considering the demand uncertainty, we constructed a rubost optimization model of the procurement production and distribution synchronization. The multigroup parallel genetic algorithm is presented to solve the model. The numerical experiment tests the practicalityof the model and the possibility of the algorithm. Analyzed the influence to the system performance of goal progrmming weightω. Study shows that robust solution is possible, but at some cost. By the proper value of weightω, we can meet the uncertain market by the less cost and little error.(4) Taking into account the uncertainty of demand production and supply, we established a robust optimization model for inventory production and transportation coordination. Multigroup parallel genetic algorithm based on simulated annealing m ethod and encoding the discrete and continue variables as a changeable genome based on binary map mode is put forward to solve the model. In numerical experiment, the result of contrastibe test shows the effectiveness of the algorithm and that by the proper value of weightω, we can meet the uncertain market by the less cost and little error.(5) Developed the prototype of robust optimization supply chain system which is made of supply chain network integration sub-system in strategic lebel, the synchronization of the plan of procurement, production and distribution plan sub-system in tactical lebel and the coordination of production inventory and transportation sub-system in operational level. The three sub-system are convergenced by custom formatdata file, each module is independent but coordinate to each other to adapt various applying. By running the example, the validity and practicality is tested and verified.