Control Systems and Computers

DOI: https://doi.org/10.15407/usim.2018.04.021

Upr. sist. maš., 2018, Issue 4 (276), pp. 21-31.

UDC 004.8

O.H. Moroz, junior research scientist, Departament for Technologies of Inductive Modelling, International Research and Training Center for Information Technologies and Systems of the NAS and MES of Ukraine, Glushkov ave., 40, Kyiv, 03187, Ukraine, Moroz.@ukr.net

Software package for inductive modeling based on combinatorial-genetic method

Introduction. The combinatorial-genetic method COMBI-GA is an effective self-organizing means of inductive modeling of complex linear and nonlinear objects, systems, and processes of various nature. A software package based on COMBI-GA is developed using the programming language MATLAB. It is intended both for solving practical modeling problems from observational data under incomplete information about an object and for investigating the capabilities of COMBI-GA, in particular, for problems of large dimensionality.

The purpose of the article is to describe the structure, interface and functionality of the software package for the inductive construction of optimal models of complex objects based on the combinatorial-genetic method.

Results. The paper presents a formal description and main steps of the hybrid algorithm COMBI-GA. The characteristics and capabilities of the software package built on the basis of this algorithm are considered in detail, particularly the general structure, working mechanism, automatic and dialog modes of the modeling process, user interface, qualitative and quantitative indicators of the simulation results etc. The analysis of the COMBI-GA effectiveness in the sense of the restoration accuracy of a given test model and the time to find it was carried out in comparison with the LASSO algorithm as well as with the sorting-out algorithms COMBI and MULTI. It is shown that all algorithms find the correct model but COMBI-GA does it much faster.

Conclusions. The developed software package provides the user with wide opportunities, in particular, it allows solving real and testing simulation problems, comparing simulation results, investigating the convergence pattern, accuracy and stability of the algorithm depending on the entered parameters to take into account the features of a specific task.

Download full text! (On Ukrainian).

Keywords: GMDH, COMBI algorithm, genetic algorithm, hybrid COMBI-GA algorithm, software package.

1. Moroz, O.G., 2016. “Sorting-Out the GMDH Algorithm with the Genetic Search of Optimal Model”. Upravlausie sistemy i masiny, 6, pp. 73-79. (In Russian).
2. Anufriev, I.E., Smirnov, AB, Smirnova, E.N., 2005. MATLAB 7. SPb .: BHV-Petersburg, 1104 p. (In Russian).
3. Stepashko, V.S., 2010. “Elements of the theory of inductive modeling”. The state and prospects of the development of computer science in Ukraine: a monograph. Col. authors Kyiv: Naukova dumka, pp. 481-496. (In Ukrainian).
4. Ivakhnenko, A.G., Stepashko, V.S., 1985. Pomekhoustoychivost’ modelirovaniya (Noise Immunity of Modelling), Kiev, Naukova dumka, 216 p. (In Russian.)
5. Moroz, O.G., Stepashko, V.S., 2016. “Comparative analysis of generators of model structures in interpolated algorithms of GMDH“. Inductive modeling of complex systems: Coll. sciences works, K .: IRTC ITS NASU, 8, pp. 133-148. (In Ukrainian).
6. Stepashko, VS, Kostenko, Yu.V., 2002. “Combinatorial selection algorithm for sequential search of optimal complexity model”. Works of the 1st International Conference on Inductive Modeling. T.1, Ch.1. Lviv,: DNDIII, pp. 72-76. (In Russian).
7. Tibshirani, R., 1996. “Regression Shrinkage and Selection via the Lasso”. Journal of the RoyalStatistical Society. Series B (Methodological), 58 (1), pp. 267-288.
   https://doi.org/10.1111/j.2517-6161.1996.tb02080.x

Received 16.11.18