(中國礦業(yè)大學(xué)采礦工程系,徐州 221008)
摘 要: 為了克服BP網(wǎng)絡(luò)自身算法的缺陷,得到更高的學(xué)習(xí)精度和更快的收斂速度,將遺傳算法、小波分析、人工神經(jīng)網(wǎng)絡(luò)和模擬退火思想結(jié)合起來,提出了一種遺傳小波網(wǎng)絡(luò):即用遺傳算法來學(xué)習(xí)小波神經(jīng)網(wǎng)絡(luò)層間的權(quán)值、尺度參數(shù)和位置參數(shù)。用兩維和三維XOR問題對其性能分別進(jìn)行了測試,取得了理想的效果;將其應(yīng)用于礦壓預(yù)報(bào),得到了比傳統(tǒng)神經(jīng)網(wǎng)絡(luò)更優(yōu)的效果。
關(guān)鍵字: 人工神經(jīng)網(wǎng)絡(luò) 遺傳小波網(wǎng)絡(luò) 礦山壓力 預(yù)報(bào)
(Department of Mining Engineering,China University of Mining and Technology, Xuzhou 221008, P. R. China)
Abstract:In order to overcome the algorithm shortcoming of BP network and obtain much higher accuracy and faster speed, a wavelet network based on genetic algorithm was put forward, which combined genetic algorithm with wavelet analysis and artificial neural network and simulated annealing thought, to study various parameters such as weights etc by genetic algorithm. Its property was tested by solving the problem of two-dimension XOR and three-dimension XOR, and good results of them were obtained. And it was also used to solve the problem of forecast in the field of ground pressure, more optimal results than typical BP network have been obtained.
Key words: artificial neural network wavelet network based on GA ground pressure forecast


