Application of Improved Operators Genetic Algorithms in Parameters Learning of Fuzzy Neural Network

摘要

The Improved Operators Genetic Algorithm (IOGA)is introduced.The original operations of crossing and variation are replaced with a simple operation of GA itself. Two parameters,i.e.cross-probability and variation-probability, are omitted here to avoid unreasonable sampling of Pc and PM and the probability reduction of the chromosome of individual with high degree of adaptability. Finally, the improved algorithm is applied to parameters learning of fuzzy neural network(FNN)by forming a FNN controller. The simulation shows that it is effective and applicable. The(IOGA)shows a strong capacity of overall optimization and provides a good solution for complex nonlinear and combinatorial optimization problem.