Partial Orders as a Basis for KBS Semantics
Partial orders are a mathematical construct currently used in denotational semantics. This construct has several properties which make it more generally applicable to knowledge-based systems (KBS) design, and in this paper we consider the role of partial orders in describing the meanings of data states in knowledge based systems. Partial orders allow formal representation of the state of information and inferences made about the external world, as stored in dynamically generated data structures of the KBS. A partial order can be augmented with a single representation of the reasoning strategies of a KBS, which includes representation of how a KBS might adapt reasoning strategies depending on the information available to it. This gives a common theoretical framework for KBS methods.
Unable to display preview. Download preview PDF.
- Lazowska, E.D., Zahorjan, J., Scott-Graham, G. and Sevcik, K.C. (1984) Quantitative System Performance., Prentice Hall, New Jersey.Google Scholar
- Leier, W. (1988) Constraint Programming Languages, Addison-Wesley.Google Scholar
- Pearl, J. (1985) How to do with probabilities what people say you can’t. C.R. Weisbin (Ed). Artificial Intelligence Applications: The engineering of Knowledge-Based Systems North-Holland 1985.Google Scholar
- Pitrat, J. (1984) An intelligent program can and must use declarative knowledge efficiently. In Elithorn, A. and Banerji, R. (Eds.) Artificial and Human Intelligence, Elsevier, North-Holland.Google Scholar
- Scott, D.S. (1976) Data types as lattices. Society of Industrial and Applied Mathematics(SIAM) journal of Computing vol 5, no 3.Google Scholar