計(jì)算機(jī)模擬系統(tǒng)-基于遺傳算法的神經(jīng)網(wǎng)絡(luò).doc
約44頁(yè)DOC格式手機(jī)打開(kāi)展開(kāi)
計(jì)算機(jī)模擬系統(tǒng)-基于遺傳算法的神經(jīng)網(wǎng)絡(luò),摘 要小波變換具有時(shí)域局部特性和變焦特性，而神經(jīng)網(wǎng)絡(luò)具有自學(xué)習(xí)、自適應(yīng)、魯棒性、容錯(cuò)性和推廣能力。把兩者的優(yōu)勢(shì)結(jié)合起來(lái)形成了小波網(wǎng)絡(luò)（wavelet neural network, wnn）。小波神經(jīng)網(wǎng)絡(luò)是由小波理論支持的一種特殊的前向控制神經(jīng)網(wǎng)絡(luò)，兼有小波變換和神經(jīng)網(wǎng)絡(luò)兩者的優(yōu)勢(shì)，是近似計(jì)算和預(yù)測(cè)領(lǐng)域廣泛流行的工具。...
內(nèi)容介紹
此文檔由會(huì)員 Facebook 發(fā)布
摘 要
小波變換具有時(shí)域局部特性和變焦特性，而神經(jīng)網(wǎng)絡(luò)具有自學(xué)習(xí)、自適應(yīng)、魯棒性、容錯(cuò)性和推廣能力。把兩者的優(yōu)勢(shì)結(jié)合起來(lái)形成了小波網(wǎng)絡(luò)（Wavelet Neural Network, WNN）。小波神經(jīng)網(wǎng)絡(luò)是由小波理論支持的一種特殊的前向控制神經(jīng)網(wǎng)絡(luò)，兼有小波變換和神經(jīng)網(wǎng)絡(luò)兩者的優(yōu)勢(shì)，是近似計(jì)算和預(yù)測(cè)領(lǐng)域廣泛流行的工具。遺傳算法和人工神經(jīng)網(wǎng)絡(luò)作為兩個(gè)工具在眾多的研究領(lǐng)域得到了廣泛的應(yīng)用，遺傳算法和神經(jīng)網(wǎng)絡(luò)本身也得到很大發(fā)展。遺傳算法體現(xiàn)了生物進(jìn)化中的四個(gè)要素，即繁殖、變異、競(jìng)爭(zhēng)和自然選擇。在本篇論文，用遺傳算法來(lái)構(gòu)建和訓(xùn)練小波神經(jīng)網(wǎng)絡(luò)，以此來(lái)近似計(jì)算和進(jìn)行預(yù)測(cè)。本文提出的遺傳算法利用分級(jí)染色體對(duì)小波神經(jīng)網(wǎng)絡(luò)的結(jié)構(gòu)和權(quán)值進(jìn)行編碼，遺傳算法聯(lián)合進(jìn)化規(guī)則來(lái)構(gòu)建和訓(xùn)練小波神經(jīng)網(wǎng)絡(luò)，同時(shí)對(duì)網(wǎng)絡(luò)進(jìn)行進(jìn)化。最后用訓(xùn)練后得到的小波神經(jīng)網(wǎng)絡(luò)用于函數(shù)近似，體現(xiàn)小波神經(jīng)網(wǎng)絡(luò)良好的近似功能。
關(guān)鍵詞 小波變換 小波神經(jīng)網(wǎng)絡(luò) 遺傳算法 函數(shù)近似
Abstract
Wavelet Neural Network, WNN）The wavelet network has been introduced as a special feed-forward neural network supported by the wavelet theory,and has become a popular tool in the approximation algorithm,which combines the wavelet theory and feed-forward neural netword.. As two kinds of tools , genetic algorithms( GA )and artificial neural network get wide applications in many research areas,and there are many variation in thenselve. Genetic algorithms shows four elements of biologic evolution : propagation , variation , competition and natural selection. In this paper, an evolutionary algorithm is proposed for constructing and training the wavelet network for approximation and forecast. This evolutionary algorithm utilises the hierarchical chromosome to encode the structure and parameters of the wavelet network, and combines a genetic algorithm and evolutionary programming to construct and train the network simultaneously through evolution.In the end, wavelet neural network after being trained is used to approximation of function to performance good approximation of function.
Keywords: Wavelet transforms;Wavelet neural network;Genetic algorithms;Approximation of function
小波變換具有時(shí)域局部特性和變焦特性，而神經(jīng)網(wǎng)絡(luò)具有自學(xué)習(xí)、自適應(yīng)、魯棒性、容錯(cuò)性和推廣能力。把兩者的優(yōu)勢(shì)結(jié)合起來(lái)形成了小波網(wǎng)絡(luò)（Wavelet Neural Network, WNN）。小波神經(jīng)網(wǎng)絡(luò)是由小波理論支持的一種特殊的前向控制神經(jīng)網(wǎng)絡(luò)，兼有小波變換和神經(jīng)網(wǎng)絡(luò)兩者的優(yōu)勢(shì)，是近似計(jì)算和預(yù)測(cè)領(lǐng)域廣泛流行的工具。遺傳算法和人工神經(jīng)網(wǎng)絡(luò)作為兩個(gè)工具在眾多的研究領(lǐng)域得到了廣泛的應(yīng)用，遺傳算法和神經(jīng)網(wǎng)絡(luò)本身也得到很大發(fā)展。遺傳算法體現(xiàn)了生物進(jìn)化中的四個(gè)要素，即繁殖、變異、競(jìng)爭(zhēng)和自然選擇。在本篇論文，用遺傳算法來(lái)構(gòu)建和訓(xùn)練小波神經(jīng)網(wǎng)絡(luò)，以此來(lái)近似計(jì)算和進(jìn)行預(yù)測(cè)。本文提出的遺傳算法利用分級(jí)染色體對(duì)小波神經(jīng)網(wǎng)絡(luò)的結(jié)構(gòu)和權(quán)值進(jìn)行編碼，遺傳算法聯(lián)合進(jìn)化規(guī)則來(lái)構(gòu)建和訓(xùn)練小波神經(jīng)網(wǎng)絡(luò)，同時(shí)對(duì)網(wǎng)絡(luò)進(jìn)行進(jìn)化。最后用訓(xùn)練后得到的小波神經(jīng)網(wǎng)絡(luò)用于函數(shù)近似，體現(xiàn)小波神經(jīng)網(wǎng)絡(luò)良好的近似功能。
關(guān)鍵詞 小波變換 小波神經(jīng)網(wǎng)絡(luò) 遺傳算法 函數(shù)近似
Abstract
Wavelet Neural Network, WNN）The wavelet network has been introduced as a special feed-forward neural network supported by the wavelet theory,and has become a popular tool in the approximation algorithm,which combines the wavelet theory and feed-forward neural netword.. As two kinds of tools , genetic algorithms( GA )and artificial neural network get wide applications in many research areas,and there are many variation in thenselve. Genetic algorithms shows four elements of biologic evolution : propagation , variation , competition and natural selection. In this paper, an evolutionary algorithm is proposed for constructing and training the wavelet network for approximation and forecast. This evolutionary algorithm utilises the hierarchical chromosome to encode the structure and parameters of the wavelet network, and combines a genetic algorithm and evolutionary programming to construct and train the network simultaneously through evolution.In the end, wavelet neural network after being trained is used to approximation of function to performance good approximation of function.
Keywords: Wavelet transforms;Wavelet neural network;Genetic algorithms;Approximation of function
