Adapting the behavior of a job-shop scheduling system

摘要

Factory scheduling consists in assigning resources (e.g. machines) and start and end times to operations. Our work is concerned with the problems of schedule generation and schedule revision when unanticipated events occur on the factory floor. SONIA is a knowledge-based scheduling system provided with a blackboard architecture for coordinating the activation of various scheduling and analyzing knowledge sources. In this paper, we focus on the various behaviors these knowledge sources can have and we gather a collection of conclusions regarding the use of various backtracking strategies and the control of constraint propagation.