CHESS: A methodology for resolving scheduling and dispatch problems in FMSs

Abstract

The objective of the research, presented herein, was to determine whether or not the makespan performance of an FMS could be improved by a methodology which draws parallels with the actions of a human chess player in resolving decision problems in an environment where optimal solutions were computationally intractable with current processing technologies. The methodology was given the acronym CHESS (Control Heuristic Embedded System Simulator) and provides a mechanism for using short-term simulation look-aheads to resolve contention problems. The research also sought to determine the relationship between performance improvement and “look-ahead distance.” The results of the research highlight the fact that the methodology could provide considerable reductions in the number of contentions and the makespan time for a given range of part types (approximately 20 percent for the test-case applications). The most pertinent results, however, are those that highlight the effectiveness of look-ahead distance in terms of improving system performance, and these may have a potential to provide a significant impact in a range of different areas.