Article ID: | iaor2002205 |
Country: | United States |
Volume: | 121 |
Issue: | 1/2 |
Start Page Number: | 211 |
End Page Number: | 250 |
Publication Date: | Aug 2000 |
Journal: | Artificial Intelligence |
Authors: | Beck J.C., Fox M.S. |
Keywords: | constraint programming |
In this paper, we expand the scope of constraint-directed scheduling techniques to deal with the case where the scheduling problem includes alternative activities. That is, not only does the scheduling problem consist of determining when an activity is to execute, but also determining which set of alternative activities is to execute at all. Such problems encompass both alternative resource problems and alternative process plan problems. We formulate a constraint-based representation of alternative activities to model problems containing such choices. We then extend existing constraint-directed scheduling heuristic commitment techniques and propagators to reason directly about the fact that an activity does not necessarily have to exist in a final schedule. Experimental results show that an algorithm using a novel texture-based heuristic commitment technique together with extended edge-finding propagators achieves the best overall performance of the techniques tested.