Social, Group and Individual Mind extracted from Rule Bases of Multiple Agents

Authors: P. Cermak

Abstract:

This paper shows possibility of extraction Social, Group and Individual Mind from Multiple Agents Rule Bases. Types those Rule bases are selected as two fuzzy systems, namely Mambdani and Takagi-Sugeno fuzzy system. Their rule bases are describing (modeling) agent behavior. Modifying of agent behavior in the time varying environment will be provided by learning fuzzyneural networks and optimization of their parameters with using genetic algorithms in development system FUZNET. Finally, extraction Social, Group and Individual Mind from Multiple Agents Rule Bases are provided by Cognitive analysis and Matching criterion.

Keywords: Mind, Multi-agent system, Cognitive analysis, Fuzzy system, Neural network, Genetic algorithm, Rule base.

Digital Object Identifier (DOI): doi.org/10.5281/zenodo.1079952

References:

[1] Takagi,T., Sugeno,M.: Fuzzy Identification of Systems and Its Application to Modelling and Control, IEEE Trans., MAC, 15, No.1, 116-132.
[2] Pokorny.M., ─îermak P.: A Non-linear Regression Analyze Through Neuro-fuzzy Modeling, The 3rd Czech-Japan Seminar on Data Analysis and Decision Making under Uncertainty, Osaka, Japan, 30-31 October, 2000.
[3] Cermak,P., Pokorny,M.: An Improvement of Non-Linear Neuro-Fuzzy Model Properties , Neural Network World, ICS AV CR, Praha, 2001, ISSN 1210-0552, pp. 503-523.
[4] Goldberg,D.E.: Genetic Algorithms in Search, Optimization, and Machine Learning, Addison-Wesley Publ. Comp., INC., 1989, ISBN 0- 201-15767-5
[5] Roupec, J.: Genetic optimization algoritms in fuzzy regulators, dissertation thesis, Brno, 1999.
[6] Cermak,P., Chmiel,P.: Parameters optimization of fuzzy-neural dynamic models, 23rd International Conference of the North American Fuzzy Information Processing Society June 27-30, 2004 Banff, AB, Canada, ISBN 0-7803-8377-X, pp. 762-767.
[7] Cermak,P.: Online Learning of Neural Takagi-Sugeno Fuzzy Model, 24rd International Conference of the North American Fuzzy Information Processing Society June 22-25, 2005 Ann Arbor, Michigan,USA, ISBN 0-7803-9188-8, IEEE Catalog No. 05TH8815C.