Detailinformation

	Diplomarbeit
Titel	Implementierung eines effizienten pattern matching Algorithmus für Produktionensysteme
Autor	Twrznik, Manfred
Erscheinungsjahr	1987
Abstract	One of the most promising branches of Computer Science is Artificial Intelligence. New programming paradigms derived from AI have already established themselves in industry. Some of the most important research areas of AI are knowledge representation and object-orientied, declarative and rule-oriented programming languages. One class of rule-oriented languages are production system languages, which are typically used for expert system development. Expericence has shown that typical production systems need more run-time for the solution of a problem than procedural programs /Barr A., Feigenbaum E.A., 1981/. This is caused by the fact that when rules are tested ("Pattern Matching") the program must iterate describes the implementation of an efficient pattern-matching algorithm which ist used in the production system language PAMELA (Pattern Matching Expert System Language). The basis of this implementation is the RETE algorithm, which has already found use in several AI languages. In this paper all implemented improvements to the RETE algorithm are described in detail and also some improvements which have not (yet) been implemented. The results of run-time comparisons with a programming language that implements the original RETE algorithm and with Prolog are also presented.
Status	Verfügbar