Negotiation-based multi-stage production control using distributed shortest path algorithm

This paper addresses a fully distributed negotiation-based model for optimally allocating jobs to be processed through multiple machines and alternative routings so that the sojourn time is minimized as confirmed jobs and alternative machines are considered. In this paper, we restrict our discussion to the case in which only one negotiation is active in the shop floor at any time. We explore issues pertaining to dynamic and fully distributed decision making (i) for deciding job completion time in a cooperative manner between machines and a job, and (ii) for choosing the best routing (having minimum sojourn time) for a job when alternative routings are available. The model is developed on the basis of a distributed shortest path algorithm using a job-initiated negotiation protocol. We prove die convergence property of the protocol to guarantee that the system agents can converge to an optimal solution in a finite time. An aggregated model for tackling alternative machines is also developed.