Article ID: | iaor19981135 |
Country: | United Kingdom |
Volume: | 5 |
Issue: | 1/2 |
Start Page Number: | 51 |
End Page Number: | 57 |
Publication Date: | Jan 1992 |
Journal: | Expert Systems with Applications |
Authors: | Pierreval H. |
Keywords: | artificial intelligence: decision support |
Priority rules are widely used in Flexible Manufacturing Systems to schedule the production in real time. No rule is globally better than any other since generally the rule performance depends on the characteristics of the system, the production objectives, and on the type of production undertaken. We are interested here in the expert system approach for the dynamic selection of the best priority rules. The main features and principles of this approach are presented. We show how to test and improve the knowledge bases using simulation and we introduce the SIGLOP system, devoted to this purpose. The benefits of both the approach and the software are illustrated through the example of a simplified flow shop. It demonstrates that an expert system can obtain better performance than a combination of priority rules. The perspectives include the fuzzy approach, simulation optimization, and multicriteria decision making.