周景宏++張文文
摘要: 為了準確控制輸電工程造價水平,提出一種基于果蠅算法優(yōu)化小波神經(jīng)網(wǎng)絡(luò)的混合預(yù)測模型。首先,對輸電工程造價影響因素進行歸一化處理,并將歸一化結(jié)果作為輸入變量;其次,利用果蠅算法對小波神經(jīng)網(wǎng)絡(luò)參數(shù)進行優(yōu)化,在此基礎(chǔ)上,利用優(yōu)化后的小波神經(jīng)網(wǎng)絡(luò)模型預(yù)測輸電工程造價;最后,將本文的預(yù)測結(jié)果和其他方法進行對比。算例結(jié)果表明,該混合模型的預(yù)測效果更理性,精度更高。
Abstract: In order to accurately control the transmission project cost, a hybrid prediction model based on the wavelet neural network optimized by the fruit fly algorithm is proposed. Firstly, the influencing factors of transmission project cost are normalized, and the normalized result is taken as input variable. Secondly, the parameters of wavelet neural network are optimized by using the fruit fly algorithm. On this basis, the optimized wavelet neural network model is used to predict the construction cost of transmission project. Finally, the forecast result of this article is compared with other methods. The results of the example show that the hybrid model is more rational and more accurate.
關(guān)鍵詞: 輸電工程;果蠅算法;小波神經(jīng)網(wǎng)絡(luò);工程造價
Key words: transmission project;fruit fly algorithm;wavelet neural network;project cost
中圖分類號:TM7;TU723.3 文獻標識碼:A 文章編號:1006-4311(2017)36-0214-02
3 結(jié)論
在輸電工程造價的預(yù)測研究中,由于影響因素較為復(fù)雜,從而使得準確的工程造價預(yù)測比較困難。本文利用果蠅優(yōu)化的小波神經(jīng)網(wǎng)絡(luò)模型進行預(yù)測,結(jié)果顯示:靜態(tài)投資工程造價的相對誤差絕對值的最大值是6.55%,最小值是5.12%,預(yù)測精度較高,符合相關(guān)誤差要求(±10%)。
參考文獻:
[1]凌云鵬,閻鵬飛,韓長占.基于BP神經(jīng)網(wǎng)絡(luò)的輸電線路工程造價預(yù)測模型[J].中國電力,2012,45(10):95-99.
[2]竇文雷,樸光綠,祝濤.輸變電工程造價控制[J].農(nóng)業(yè)科技與裝備,2012,32(11):51-52.
[3]熊瑋.輸電工程造價估算的混合模型研究[D].北京:華北電力大學(xué),2016:32-45.
[4]Niu Dongxiao. Research on neural network prediction of power transmission and transformation project cost based on GA-RBF and PSO-RBF [J].Applied Mechanics and Materials,2014(6):231-234.
[5]王佼.應(yīng)用灰關(guān)聯(lián)分析的PSO-SVR工程造價預(yù)測模型[J].華僑大學(xué)學(xué)報(自然科學(xué)版),2016,37(6):708-713.
[6]彭光金.小樣本工程造價數(shù)據(jù)的智能學(xué)習(xí)方法及其在輸變電工程中的應(yīng)用研究[D].重慶:重慶大學(xué),2010:45-68.
[7]韋俊濤.電力工程造價小樣本估算模型研究[D].重慶:重慶大學(xué),2009:61-71.
[8]Ding Shifei.An optimizing method of RBF neural network based on geneticalgorithm [J].Neural Computing and Application,2012(2):333-336.
[9]Chen Y, Peter B L, Guan C, et al. Short-term load Forecasting: similar day-based wavelet neural networks[J].IEEE Transactions on Power Systems, 2010,25(1):322-330.
[10]Guan C, Peter B L, Michel L D, et al. Very short-term load forecasting: wavelet neural networks with Data Pre-Filtering [J].IEEE Transactions on Power Systems,2013,28(1):30-41.
[11]侯逸文,沈炯,李益國.基于小波神經(jīng)網(wǎng)絡(luò)的火電單元機組負荷系統(tǒng)建模仿真研究[J].中國電機工程學(xué)報,2003,23(10):220-224.endprint