A two-stage evolutionary process for designing TSK fuzzy rule-based systems
- 1 January 1999
- journal article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Systems, Man, and Cybernetics, Part B (Cybernetics)
- Vol. 29 (6), 703-715
- https://doi.org/10.1109/3477.809026
Abstract
Nowadays, fuzzy rule-based systems are successfully applied to many different real-world problems. Unfortunately, relatively few well-structured methodologies exist for designing and, in many cases, human experts are not able to express the knowledge needed to solve the problem in the form of fuzzy rules. Takagi-Sugeno-Kang (TSK) fuzzy rule-based systems were enunciated in order to solve this design problem because they are usually identified using numerical data. In this paper we present a two-stage evolutionary process for designing TSK fuzzy rule-based systems from examples combining a generation stage based on a (mu, lambda)-evolution strategy, in which the fuzzy rules with different consequents compete among themselves to form part of a preliminary knowledge base, and a refinement stage in which both the antecedent and consequent parts of the fuzzy rules in this previous knowledge base are adapted by a hybrid evolutionary process composed of a genetic algorithm and an evolution strategy to obtain the final Knowledge base whose rules cooperate in the best possible way. Some aspects make this process different from others proposed until now: the design problem is addressed in two different stages, the use of an angular coding of the consequent parameters that allows us to search across the whole space of possible solutions, and the use of the available knowledge about the system under identification to generate the initial populations of the Evolutionary Algorithms that causes the search process to obtain good solutions more quickly. The performance of the method proposed is shown by solving two different problems: the fuzzy modeling of some three-dimensional surfaces and the computing of the maintenance costs of electrical medium line in Spanish towns. Results obtained are compared with other kind of techniques, evolutionary learning processes to design TSK and Mamdani-type fuzzy rule-based systems in the first case, and classical regression and neural modeling in the second.Keywords
This publication has 19 references indexed in Scilit:
- A learning process for fuzzy control rules using genetic algorithmsFuzzy Sets and Systems, 1998
- A three-stage evolutionary process for learning descriptive and approximate fuzzy-logic-controller knowledge bases from examplesInternational Journal of Approximate Reasoning, 1997
- Fuzzy connectives based crossover operators to model genetic algorithms population diversityFuzzy Sets and Systems, 1997
- A CLASSIFIED REVIEW ON THE COMBINATION FUZZY LOGIC–GENETIC ALGORITHMS BIBLIOGRAPHY: 1989–1995Published by World Scientific Pub Co Pte Ltd ,1997
- Applicability of the fuzzy operators in the design of fuzzy logic controllersFuzzy Sets and Systems, 1997
- Tuning fuzzy logic controllers by genetic algorithmsInternational Journal of Approximate Reasoning, 1995
- Improving a Fuzzy Inference System by Means of Evolution StrategyPublished by Springer Science and Business Media LLC ,1994
- ANFIS: adaptive-network-based fuzzy inference systemIEEE Transactions on Systems, Man, and Cybernetics, 1993
- Generating fuzzy rules by learning from examplesIEEE Transactions on Systems, Man, and Cybernetics, 1992
- Fuzzy setsInformation and Control, 1965