Regular Memory Structures and Operation Domains of Intelligent Systems

Abstract

The concept of a regular memory area for an intelligent system (IS) is considered. The formalized description of the memory of a separate IS component is based on an infinite saturated binary tree. Knowledge is stored in special memory subareas in the form of semantic hierarchies. This knowledge constitutes the memory content, represented by its semantic structure. This structure integrates knowledge generated and transformed by morphisms and evolutions of knowledge, which, in turn, are used to implement the IS goals. The system of knowledge morphisms and knowledge evolution is used for IS modeling. This system allows one to describe the areas of initial data and values for these morphisms and evolution using regular expressions. The family of these sets generalizes the system of classes of morphism domains. These classes are developed for the knowledge formalisms and knowledge processing flowcharts used. The application of regular expressions to describe memory structures of IS components makes it possible to construct high-level mathematical models for big and complex intelligent systems. These models allow one to develop distributed memory control schemes for knowledge processing flows and processes in intelligent systems.