A window time negotiation approach at the scheduling level inside supply chains

International audienceWe consider a supply chain consisting in a network of enterprises with independent decision centers. The products of the assembly enterprise are produced using components and/or sub products supplied by other enterprises or external suppliers. We are at the scheduling level and each enterprise builds its own schedules, which are dependent. Negotiations between decision centers can be expressed in term of penalty functions associated with soft and hard release and due dates. At each negotiation point, hard release and due dates are considered as imperative constraints, while soft release and due dates define soft intervals and induce earliness and tardiness penalties. Any interval can be modified during the negotiation process. A global solution is searched by an iterative decomposition approach including alternatively bilateral negotiations of the soft and hard constraints and just in time scheduling, minimizing the local total sum of penalties, built by approximation approaches. We assume that production centers are flow shop. To solve each local just-in-time scheduling problem, we propose an approximation approach based on meta-heuristics, in which a solution is described by the job order on each machine and is evaluated using a “pert cost” algorithm. A semi-decentralized control is suggested to assume the negotiation convergence