Abstract
This paper discusses the application of artificial intelligence (AI) techniques for solving production management problems. Over the past years companies have been forced to alter their overall management strategies radically. Now, speed, flexibility, responsiveness, adaptiveness, and quality have become increasingly important. Companies must have the ability to respond to sudden market changes; and in order to meet these new requirements, the company has to master new technologies and new organizational forms. At all levels, organizations have to deal with quick and high quality decisions, and this creates the need for not only conventional support systems, but also systems that are able to advise the decision makers in performing complex and knowledge-intensive tasks.
In particular, industrial scheduling within the production management (PM) domain is addressed. In general, industrial scheduling using analytical techniques is an NP-complete problem. Scheduling problems have been the target of a substantial amount of research within the operations research (OR) community, yet conventional OR techniques applied to scheduling have had little impact in industry. Analytical results have been achieved only for problems of moderate complexity.
Using AI techniques the inherent complexity of the scheduling problem can be reduced to a tractable size. We present a methodology for synthesizing detailed production schedules, based on the state-space search paradigm within AI. The paradigm is extended to include heuristic pruning, giving the ability to utilize the company-specific production strategies and knowledge employed in manual planning. Finally, we present a generic, industrial scheduling system, PLATO-PS, using these techniques.
In particular, industrial scheduling within the production management (PM) domain is addressed. In general, industrial scheduling using analytical techniques is an NP-complete problem. Scheduling problems have been the target of a substantial amount of research within the operations research (OR) community, yet conventional OR techniques applied to scheduling have had little impact in industry. Analytical results have been achieved only for problems of moderate complexity.
Using AI techniques the inherent complexity of the scheduling problem can be reduced to a tractable size. We present a methodology for synthesizing detailed production schedules, based on the state-space search paradigm within AI. The paradigm is extended to include heuristic pruning, giving the ability to utilize the company-specific production strategies and knowledge employed in manual planning. Finally, we present a generic, industrial scheduling system, PLATO-PS, using these techniques.
