Resolution of constraint inconsistency with the aim to provide support in anaesthesia
We have investigated the possibilities to use constraints to provide decision support in Anaesthesia. It became apparent that concepts are required to deal with inconsistency due to three causes: differences between defaults and specific information, imprecision in measurements or knowledge, and incompatibility of different preferences. Differences between defaults and specific knowledge can be resolved by attaching priorities to constraints. Imprecision in measurements or in knowledge can be resolved by using intervals instead of point values. For resolution of inconsistency due to incompatibility of preferences, next to priorities and intervals, relaxation is required. Thus, the constraint satisfaction system must support three concepts for inconsistency resolution: priorities, intervals and relaxation.
For constraint-satisfaction a number of methods has been developped: local propagation, relaxation, constraint propagation, backtracking, and symbolic rewriting. For each of these methods we investigated which of the concepts required for inconsistency resolution they can support. Furthermore it was investigated how different constraint satisfaction methods can be combined such that two or all three of the forementioned concepts can be supported.
KeywordsConstraint Satisfaction Expert Systems Architecture Medical Information Systems
Unable to display preview. Download preview PDF.
- Alan Borning. Thinglab: A Constraint-Based Simulation Laboratory. PhD thesis, Computer Science Department, Stanford University, 1979.Google Scholar
- Alan Borning, Robert Duisberg, Bjorn Freeman-Benson, Axel Kramer, and Michael Woolf. Constraint hierarchies. In Norman Meyrowitz, editor, Proc. OOPSLA-87, pages 48–60, New York: NY, 1987. ACM Press.Google Scholar
- S. van Denneheuvel, K.L. Kwast, P. van Emde Boas, F. de Geus, and E. Rotterdam. Symbolic computation in RL/1. In Proc. DEXA-91, pages 559–564, Berlin, 1991. Springer Verlag.Google Scholar
- Sieger van Denneheuvel. Constraint Solving on Database Systems; University of Amsterdam, department of Mathematics and Information Science, October 1991.Google Scholar
- A. Fred de Geus and Ernest P. Rotterdam. Decision Support in Anaesthesia. PhD thesis, State University of Groningen, Faculty of Medicine, The Netherlands, December 1992.Google Scholar
- Wm. Leler. Constraint Programming Languages; Their Specification and Generation. Addison Wesley, 1988.Google Scholar
- A.K. Mackworth. Constraint satisfaction. In Stuart C. Shapiro, editor, Encyclopedia of AI, pages 205–211. John Wiley & Sons, 1987.Google Scholar
- Peter van Emde Boas. RL, a language for enhanced rule based database processing, working document. RJ 4869 (51299), IBM Research Laboratory, San Jose: CA 95193, 1985.Google Scholar
- Peter van Emde Boas. A semantical model for integration and modularization of rules. In Proc. MFCS 12 Bratislava, pages 78–92, Berlin, August 1986. Springer Verlag.Google Scholar
- Pascal Van Hentenryck. Constraint Satisfaction in Logic Programming. MIT Press, Cambridge: MA, 1989.Google Scholar