FUNCTIONAL-SEMANTIC NETS — THE UNIVERSAL FORMALISM FOR DEFINING, DESIGNING AND ESTIMATING THE QUALITY OF FUNCTIONING OF MAN-MACHINE SYSTEMS

Abstract

A new approach is described to defining, designing and estimating the quality of discrete processes of man-machine systems (MMS) functioning. The functioning model MMS as functional-semantic networks is presented by two levels: 1) external operations, by functional networks (a stochastic model similar to PERT but taking into account not only time but also the probabilities of possible random outcomes — man's mistakes and machine faults); 2) internal (planning-mental) operations, by semantic networks realizing control over functional networks by means of the methods of artificial intellect (PROLOG programmes). Functional networks have quantitative characteristics of the quality of modelled process: time, reliability, and cost. Analizer of functional networks PROSTAK automatizes the estimation process and allows us to realize a computer-aided design of MMS and the synthesis of the functioning and control algorithms. The apparatus of functional-semantic networks is universal in terms that it can be used to describe, design and estimate the quality of functioning of MMS for various application: manufactural, technological, information-control management, transportational, etc. In this case man's operations are described by unified formalisms, which allows us to estimate variants of MMS organization with different level of mechanization and automatization. Functional-semantic networks unite both the procedural component of model MMS (as a functional network) and the declarative component (as a semantic network). This leads to the appearance of a fundamentally new class of MMS-models. In terms of mathematics, functional — semantic networks provide with new possibilities of modelling, estimating and solving applied problems in ergonomics, the theory of MMS reliability, the theory of MMS control, the theory of MMS technology and organization.