Bid construction scheme for job flow time reduction in auction-based fully-distributed manufacturing systems

Abstract

This study presents a novel bid construction scheme for an auction-based fully-distributed manufacturing system. The scheme is applied to the planning and control of manufacturing activities. A set of independent machining resources in the system, each with several ongoing set-up dependent jobs, compete for an upcoming job. The manner in which a resource constructs a bid determines job allocation and resource utilization and, consequently, job flow times. Resources that employ simple heuristic methods (e.g., based on a first-come-first-served scheme) to construct bids, often neglecting setup considerations, limit their applicability and performance quality. Here, the bid construction problem of a resource is formulated as a setup-sequencing (SS) problem. The optimal sequencing theory is then applied to develop a dynamic programming (DP)-based bid construction algorithm. The proposed algorithm has a computational complexity lower than that of the best-known algorithm for the SS problem. Additionally, parallel and distributed computing technology is utilized to model and implement this auction-based manufacturing control system in order to gain further insights into the effectiveness of the proposed DP-based bid construction scheme. Experimental results demonstrate that the proposed bid construction scheme performs well, and its contribution becomes obvious in the case of larger setup times and lot sizes.