INTELLIGENT SYSTEMS

Optimization in expert systems for the sequential selection of objects based on user surveys as part of natural language information agents

A. A. Sokolov

Dorodnitsyn Computing Centre of the Russian Academy of Sciences, Moscow

Abstract: The construction of expert systems with serial poll of users which search for objects in the knowledge base of this ES is considered. The problem of minimization of an average of steps of procedure of search is posed. It is shown that for minimization of duration of procedure the data should be introduced to the knowledge base not only about a subject domain, but also about users target groups. For use ES in natural language information systems (Infs) the description of the knowledge base by means of indistinct sets and linguistic variables is imposed. Estimations and quasioptimal algorithms of the decision of a task in view are offered. Introduction examples are given.

Keywords: экспертная система, минимизация, естественно-языковая информационная система, нечеткие множества.

UDC: 519.872.7

Received: 14.06.2009
Revised: 28.06.2009

Bibliographic databases:

•