A knowledge based framework for developing and customizing schedulers

Abstract

The paper describes a general purpose scheduler into which one can add knowledge necessary to solve specific scheduler problems. The approach typifies how software should be developed for specific classes of applications. Using a general scheduler is analogous to the use of an expert system shell to which one adds application specific rules. The authors performed experiments for adding knowledge to three classes of schedulers-job shop schedulers, transportation schedulers and time tabling schedulers. They show how the framework provides flexibility in performing experiments.>