Conception de systemes manufacturiers cellulaires dans un environnement de chaines d'approvisionnement

Classical cellular manufacturing system (CMS) design are performed on a single plant and are no more appropriate in an environment where material procurement and customer delivery processes issues with the selection of multi-plant production system locations have to be considered simultaneously. Integration of these issues is a crucial challenge to supply chain managers.;This thesis deals with the multi-plant CMS design in a supply chain environment. This context is established with considering the supply or the customer delivery processes, and with linked manufacturing plants. Three problems are examined.;The first problem considers linked multi-plant CMS design integrated to material supply process. The solution approach is based on a proposed linearised model where the total system design cost to minimise combines the cost of the cellular configuration set on existing plants and the costs linked to supplier process selection. Experimentation demonstrates the potential benefits gained through increasing routing flexibility over plants on investment costs and the effect of integrating the supply process in a multi-plant cellular manufacturing configuration.;The second problem combines multi-plant CMS design to manufacturing plant selection and customer demand allocation decisions. We propose a mathematical model and develop a simulated annealing based approach which aims to select the manufacturing plants to open and define the machine cells locations and structures and assign customer’s demand. We demonstrate the efficiency of the solution approach and show the cost savings due to integrated multi-stage multi-plant CMS with customer allocation decisions, compared to a sequential decision process.;The third problem examines a context of multi-period customer demand where multi-plant production planning decisions, part transfer between plants, system reconfiguration and dynamic customer allocation decisions are coupled with dynamic multi-plant CMS configuration. It is shown that considering these issues in an integrated model reduces total design cost and enhances manufacturing and delivery flexibility.