A Hybrid GA-BP Self-Learning Algorithm for Fury Logic System
模糊逻辑系统的GA+BP混合学习算法

A new algorithm, in which back propagation (BP) algorithms of various learning rate are used to produce a little parts of population for genetic algorithm (GA), is introduced for the self learning of fuzzy logic system. In particular, the global convergent characteristic of the genetic algorithm is used to find the possible extremums in the whole area, and the great feature of the BP algorithm, that is, error descend in the direction of grads, is used to fast search about the extremums. Thus, the better combination of global convergence and fast search is obtained. Simulation results show that the hybrid algorithm is far more efficient not only than the GA but also than the back BP.