Planification d'une chaîne logistique: approche par satisfaction de contraintes dynamiques

This work focuses on the supply chain operational and tactical planning problem in an uncertain environment. We propose a tactical planning model, based on a two-level decisional structure. Adopting a dynamic process, which enables data updating at each planning step, the first level performs a plan by searching for the best compromise between available decision-making levers for production costs and capacity. The second level establishes an aggregate scheduling of production tasks by minimizing the total weighted completion time. The use of an integrated decision structure involves interaction between the upper and lower decision levels constraints, implemented by constraints known as energy conservation constraints. Our approach is formulated according to a constraint satisfaction problem (CSP) and evaluated by simulation under uncertain data. For this, we have developed various experiments related to the variation of customer demand, resource capacity, and the re-planning of demand. All the experiments are carried out by two different methods: a method based on a static CSP and a method based on a dynamic CSP. The performance of a scheduling/planning solution is reported through a set of robustness and stability measurements. Results of experiments confirm the performance of the method based on a dynamic CSP.