Name
Abstract | ||
|---|---|---|
One of the challenges in process control is providing reliable control of poorly understood systems. Before such a system can be controlled we must first be able to predict its future behavior-so that we know what control action is necessary. This paper presents two approaches to this prediction task, both using qualitative models augmented by records of historical system behavior. Our hypothesis is that qualitative information about a system is more easily available than quantitative equations; moreover, the information need not be complete or totally correct. We restructure the historical information into a case-base suitable for the prediction task, and use the qualitative model to identify the attributes to use as case-indices. The case-base then provides the quantitative information needed for the prediction task. Our techniques are extensively evaluated on data taken from a real-world system. |
Year | Venue | Keywords |
|---|---|---|
1995 | IJCAI | process control,information need |
Field | DocType | ISBN |
Information needs,Computer science,Work in process,Artificial intelligence,Machine learning,Qualitative reasoning | Conference | 1-55860-363-8 |
Citations | PageRank | References |
0 | 0.34 | 9 |
Authors | ||
3 | ||
Name | Order | Citations | PageRank |
|---|---|---|---|
| Bradley L. Richards | 1 | 209 | 37.47 |
| Boi Faltings | 2 | 3586 | 331.33 |
| Peter Duxbury-Smith | 3 | 0 | 0.34 |