Research On Robust Optimization Model Of Supply Chain Based On Risk Control

The presence of a variety of uncertainties is the main source of supply chain risk, so thatit is an important issue of the supply chain risk management to implement effective supplychain risk control. However, different risk factors always have different probability ofoccurrence and therefore have different impact, thus, this thesis selects the demanduncertainty which is most common and also with the most extensive influence as the researchbackground. By the method of robust optimization, the risk-sharing and risk appetite controlproblems in the supply chain risk management will be considered in this thesis throughestablishing corresponding robust optimization model and the analysis of the related problems.Specific studies are as follows:1. This thesis introduces the relevant research of supply chain risk control, and reviewsthe development and application of robust optimization. A risk-sharing strategy for thedecentralized supply chain and a risk appetite control strategy for the integrated supply chainare put forward through qualitative and quantitative research on supply chain risk.Simultaneously, it introduces the basic idea of robust optimization method and describes theadvantages and scope of application of the various robust optimization models.2. A linear robust optimization problem subjected to demand uncertainty is consideredwhen taking into account the quantity flexibility contract which can achieve risk sharingamong supply chain members in proportion. Trough the analysis of the historical data, afluctuation range of demand for commodity can be determined. Given the nice performanceof robust optimization in handling problems with uncertainties, and the advantages of quantityflexibility contract with respect to risk-sharing mechanism, this thesis studies a two-stagesupply chain system, establishes a robust operation model based on the quantity flexibilitycontract, using the stackelberg game theory, gives the robust strategies and discusses theinfluence on the supply chain members’and the overall revenue.3. Considering the risk preference of the integrated manager, we study the dynamicproduction and inventory management problem for the purpose of risk preference control. Asupply chain framework with I producer enterprises, one warehouse and one waste treatmentbase is constructed to describe the T-stage production and inventory management problem ,then we get one joint ellipsoid uncertainty set when taking the risk preference of decisionmakers into account, and set up one uncertain optimization model with the pursuit ofmaximizing the overall revenue. By using the robust idea, we translate the uncertainoptimization model into one deterministic linear robust counterpart, and discusses therobustness of the solution corresponding to the reliability and validity of the model.4. For the plans made by the decision makers are often closely related to the demands ofprior periods in the dynamic production management issues, further, an affine adjustablerobust optimization problem is resolved. Considering the advantage of affinely adjustablerobust counterpart method in dealing with problems with uncertainties in the dynamic context,we develop an uncertain optimization model in pursuit of maximizing the overall revenue through adaptively controlling multi-period production policies, and equivalently convert it toone deterministic robust counterpart which is in fact a tractable second order cone problem.The solutions or approximate solutions of above two linear robust optimization modelsand the affine adjustable robust optimization model can be obtained by translating thesemodels into the corresponding deterministic mathematical programming which arecomputationally tractable by Lingo, Matlab, Cplex, GAMS and some other commonlycommercial software, so that the actual decision makers only need to focus on the modelconstruction with neglect of the model solution algorithm design. Finally, according to twokind of linear robust optimization model, the numerical example is analyzed to show theeffectiveness of robust optimization method and describe the influence of robust strategies onthe supply chain risk control objectives.