Fuzzy Wavelet Neural Networks Control And Its Application Based On Ant System Algorithm

The characteristic of ant system algorithm that is a new intelligence optimized algorithms, is studied in this paper firstly and used to design fuzzy wavelet neural network optimally. Then this paper presents a fuzzy wavelet neural network model reference adaptive inverse control strategy based on ant system algorithms, also according to the characteristic and control requirements of complex non-linear plants (the EFAT/P series process control experimental device), this paper devises two different experimental schemes to realize real-time control, which both obtain the anticipated result.Main research works in this paper are as follows:1. Combination of wavelet function and fuzzy neural network. Considering the subject function deficiencies of language variable layer in the existing fuzzy neural network, fuzzy wavelet neural network is put forward using the good dilation and translation characteristic of wavelet transformation.2. Determination of control strategy. On the basis of comparing and analyzing several representative neural network controls models, aiming at the characteristics and control requirement for complex non-linear plant, a model reference adaptive inverse control method based on fuzzy wavelet neural network is suggested.3. Setting up the neural network study algorithm. On the basis of analyzing the optimal performance of ant system algorithm and the deficiency of optimize performance in existing ant system algorithms, hybrid adaptive ant system algorithm with pheromone weight multiplier of normal distribution and pheromone balance operator of uniformity distributed is put forward, and is applied to parameter and structure optimal design of the fuzzy wavelet neural network;4. Aiming at the characteristic of complex non-linear plant, integrative processcontrol systems containing industrial process controller, intelligent apparatus and YANHUA facility with MCGS configuration software as platform, is used to carry on real-time control. Experimental results show the validity and feasibility of the control strategy;5. On the basis of experiment, 16-bit single-chip computers process control system based on Intel 80196KC chip is designed. Debugging result indicates that the control strategy presented in this paper can not only enhance the stability and speediness, but also improve the self-learning and anti-jamming ability of control system...