Research On Optimization Model And Its Algorithm In Supply Chain Management Under Uncertainty

Supply chain management (SCM) emphasize the optimization of all processes and the improvement of the whole system performance. Considering the uncertainties, the decision-maker needs to find optimal solution for achieving the objective in constrained or restricted resource circumstance. This work is generally finished by optimization model and algorithm.The paper analyzes domestic and foreign research status about supply chain management in operation research firstly. Next it points out the uncertainties and their influence in supply chain, expatiates the idea on supply chain management modeling and optimization, and reviews robust optimization and stochastic programming in uncertain optimization theory. Finally, several crucial problems are studied in supply chain management under certainties. These problems include how to design integrated supply chain network in strategic level, how to synchronize procurement production and distribution plan in tactical level, how to coordinate production inventory and transportation in operational level. The mathematic model is constructed for them and the algorithm is presented respectively. The model application and algorithm effectiveness are validated by contrastive test in the computational sample.The main creative points in the paper include the below aspects.1. Considering uncertainties of the parameters, a robust optimization model is constructed for supply chain network design, which integrates supplier selection problem with facility decision problem. The feasibility is analyzed for solving the mode. Combining tabu search technique and all or nothing principle, the heuristic approach is put forward to solve the model. In the computational example, the result of contrastive test validates that the performance of algorithm is outstanding on convergence and computational time. On the other hand, and also shows the robust optimization model can reduce business risks effectively when it is applied to design supply chain network.2.Taking account into demand uncertainty, a two-stage stochastic programming model with recourse problem is established for synchronized procurement production and distribution plan. By Monte Carlo simulation of stochastic demand, the hybrid simulation-optimization policy is brought forwardto solve the model as well as the principle to select size of the sample. Benders decomposition algorithm and its accelerated technique are presented to solve the deterministic equivalent of primal problem. In the computational example, the result of contrastive test shows the performance of algorithm is excellent on solving large-scale problem, and also makes out the plan can save the cost when the stochastic programming model taking into account demand uncertainty determines it.3.Focusing on uncertainty of the material supply, the production process and the customer demand, a stochastic programming model with probability constraint is presented to coordinate production, inventory and transportation process in supply chain. The equivalent transformation approach is put forward for the probability constraint. By relaxation of the constraint, the primal problem is decomposed into production subproblem, inventory subproblem and vehicle schedule subproblem. Then the solving policy is proposed for these subproblems. The sub-gradient algorithm is designed to solve lagrangian dual problem according to its characteristic. In computational example, the result of contrastive test shows the algorithm effectiveness. In addition, it reflects that the coordinated production inventory transportation can reduce the total cost in the system.4. Made use of algebraic modeling language (AML) LINGO, the mathematic model is programmed. Via dynamic link library program, the large-scale uncertain optimization problem is solved by the combination of simulation program, heuristic approach composed high-level language and accurate algorithm based on mathematic programming.