Flexibility Design For Production System:Supply Chain And Workshop

This dissertation focuses on supply chain flexibility design and performance analysis.With the development of the economy,market is becoming more uncertain than before.Supply chain flexibility is proposed as a hedge to these uncertainties.This dissertation studies supply chain flexibility from two levels: strategic and operational levels.For the strategic level,design problems are studied: the process flexibility design problem in a single-stage and single-period supply chain,and the flexibility design problem in a multistage and multi-period supply chain.A two-stage stochastic programming model and a dynamic programming model are proposed for the two design problems,respectively.For the operational level,we propose a flexible rescheduling problem based on the perspective of human behaviors and solve it with mixed integer programming methods.Process flexibility design problem in single stage and single period supply chains is common in industry.Different from those studies with specific settings in the existing research,this dissertation studies the problem based on more general settings.Specifically,the study has no assumptions of balanced,symmetric systems with homogeneous cost parameters.A two-stage stochastic programming model is proposed and the adapted Lshaped method is applied to solve the problem.Several enhancements aiming to improve the L-shaped method's effectiveness are discussed.Comparing results with existing heuristic design methods,the method proposed in this dissertation can be applied to more general situations.Additionally,the proposed method can be easily adapted to the problem of considering capacity investment and flexibility design simultaneously.A dynamic programming model is proposed to study the flexibility design problem in multi-stage and multi-period supply chains.The model can be decomposed into subproblems dependent on stages and products.Furthermore,based on the decomposed sub-problems,this dissertation proves that the optimal policy for lower echelons is basestock policy,which greatly simplifies the process of solving the original model.Based on these results,the dissertation analyzes the impacts of introducing multiple stages and multiple periods on the performance of long chain flexibility.The study shows that the long chain flexibility still performs almost as well as total flexibility.As for the operational level problem in workshops,this dissertation studies the flexible rescheduling problem considering disruptions and proposes a mixed integer programming model.The model aims to minimize decision makers' negative deviation from the original schedule,which is quite different from those classical rescheduling studies.Based on the comparison of two rescheduling policies(i.e.,as-soon-as-possible policy and as-late-aspossible policy)under the proposed model and the classic model,we find that the newly proposed model prefers to handle disruptions as soon as possible,which would be helpful to alleviate the impacts of disruptions on supply chain flexibility.Based on these studies,this dissertation provides not only academic contribution to flexibility studies but also managerial insights to practical supply chain management.