Study On Designing Resilient Supply Chain Networks By Robust Optimization Methods

Supply chain network is one of the hot issues in current research.With the rapid development of economy and the enhancement of the influence of economic globalization,enterprises often encounter some unforeseen disruption events,which more or less have negative impacts on their supply chains.Therefore,it is of great significance to construct a resilient supply chain.In the context of social uncertainty,many decision makers are more concerned about the sources of uncertainty in the supply chain construction process,such as demand,unit transportation cost,etc.Decision makers should not only focus on the total costs of the resilient supply chain,but also consider how to withstand the risk of disruption.Therefore,in order to make the supply chain have the ability to resist the disruption risk,this thesis studies the connectivity of the performance of the resilient supply chain network and puts forward a robust optimization model and two distributionally robust optimization models for resilient supply chain network construction.The main contents of this thesis include the following four aspects:(1)In the context of uncertainty,the resilient supply chain network is constructed.The resilient objective is described by the product of the number of node-disjoint paths and demand for each retailer that can represent the connectivity.The multi-objective robust optimization model and distributionally robust optimization models based on resilience and cost are proposed.Using the -constraint method,the two goals of resilience and cost are transformed into a single goal;(2)As for the uncertain demand of retailers,the uncertainty set is used to describe the uncertain demand without knowing its distribution information.Because the established model is a semi-infinite programming model,which is difficult to be solved directly.For this purpose,we use the robust optimization method,Based on the box-ellipsoid perturbation set and duality theory,we transform it into a computable form and apply it to an illustrative example to illustrate the rationality of robust multi-objective resilient supply chain network model;(3)As retailers' demands and unit transportation cost are uncertain,under the condition of obtaining partial distribution information,they are described by the ambiguity set and uncertainty set.Under the box-ball-budgeted perturbation set,the distributionally robust resilient supply chain network models are constructed.The proposed models are semi-infinite models with ambiguous chance constraints,which cannot be directly calculated and solved.The safe convex approximation forms are obtained by the safety approximation related lemma;(4)Applying the proposed model to a real case of fresh vegetable supply chain distributed by a group of potato producers in Italy to several retailers in Europe,and carrying out comparative experiments and sensitivity analysis.The new ideas in this thesis are summarized in three aspects:(1)In the aspect of model building,the concept of node-disjoint path is introduced,and a new robust multi-objective resilient supply chain model and two distributionally robust resilient supply chain models are proposed;(2)In the aspect of model analysis and solution,because the proposed model can not be solved directly,the new models are treated by the robust optimization method.The computable form of the new models are proved;(3)In the aspect of the innovative application of the models,a real potato supply chain is taken as an example,and the experimental results can illustrate the practicability of the proposed models and the rationality of the robust method.