PAIS Lab Seminar: Structural and parametric adaptive synthesis of discrete systems with a given behavior

At the session of our seminar, Petrosov David Aregovich, PhD, Head of the Department of Computer Science and Information Technology Belgorod State Agricultural University after V.Y. Gorin delivered the report «Structural and parametric adaptive synthesis of discrete systems with a given behavior»

It is proposed to use combined approach with modelling and simulation techniques, bioinspired evolutionary modelling and theory of Petri nets for the construction of discrete systems with given behavior.

Petri nets are used for describing of functioning and modelling of the adaptive structurally-parametric synthesis of synthesized system.

The evolutionary approach is used to solve optimization problems that occurs during systems synthesising. Model description of genetic algorithm for desired structure searching and discrete object parameters are based on the nested Petri nets. It is allowed herewith to vary component composition of the system, operating parameters and initial states of components and intercomponent links (that are grouped in a separate layer as "intercomponent bus"). There are the selected "points" in genetic algorithm, which improves process of structurally-parametric synthesis of discrete systems with a given behavior by changing it’s operation modes. It is proposed to use neural networks as control mechanism for structurally-parametric synthesizing process. Simulation of neural network RTRN class (Real Time Recurrent Network) is carried out with the help of informational Petri nets.

Computational experiments for approbation of the proposed combined approach for solving problem of the adaptive structurally-parametric synthesis of discrete systems with given behavior are carried out with using of the parallel computing tools