计算机工程与应用
計算機工程與應用
계산궤공정여응용
COMPUTER ENGINEERING AND APPLICATIONS
2010年
11期
42-45
,共4页
孙明%赵琳%徐耀群%丁继成
孫明%趙琳%徐耀群%丁繼成
손명%조림%서요군%정계성
Gauss小波%混沌神经网络%优化问题
Gauss小波%混沌神經網絡%優化問題
Gauss소파%혼돈신경망락%우화문제
Gauss wavelet%chaotic neural network%optimization problems
基于神经元的自反馈项可产生混沌的现象,将Gauss小波函数作为混沌神经元的自反馈项.分析了Gauss小波的尺度和平移参数对神经元动力学的影响,提出了自反馈连接权和Gauss小波尺度双重模拟退火的混沌神经元.构建了混沌神经网络模型.分析了由Ganss小波函数产生的附加能量函数对网络优化能力的影响.优化问题的仿真结果表明,该网络能够以较快的速度找到优化问题的全局最优解.
基于神經元的自反饋項可產生混沌的現象,將Gauss小波函數作為混沌神經元的自反饋項.分析瞭Gauss小波的呎度和平移參數對神經元動力學的影響,提齣瞭自反饋連接權和Gauss小波呎度雙重模擬退火的混沌神經元.構建瞭混沌神經網絡模型.分析瞭由Ganss小波函數產生的附加能量函數對網絡優化能力的影響.優化問題的倣真結果錶明,該網絡能夠以較快的速度找到優化問題的全跼最優解.
기우신경원적자반궤항가산생혼돈적현상,장Gauss소파함수작위혼돈신경원적자반궤항.분석료Gauss소파적척도화평이삼수대신경원동역학적영향,제출료자반궤련접권화Gauss소파척도쌍중모의퇴화적혼돈신경원.구건료혼돈신경망락모형.분석료유Ganss소파함수산생적부가능량함수대망락우화능력적영향.우화문제적방진결과표명,해망락능구이교쾌적속도조도우화문제적전국최우해.
Based on the phenomena that the neuron's self-feedback can generate chaos, Gauss wavelet function is used as the self-feedback of chaotic neuron in this paper.The effect of the scale and shift parameters in Gauss wavelet on the neuron dynamics is analyzed and the chaotic neuron with double simulated annealing mechanisms realized respectively by the self-feedback connection weight and the Gauss wavelet scale is proposed.The chaotic neural network is constructed and the effect of the additional energy function generated by Gauss wavelet function on the optimization ability of the network is studied.The simulation results on optimization problems show that the proposed network can find the global optimum of the optimization problems with a faster speed.