本文關(guān)鍵詞： 可靠性分析 參數(shù)估計(jì) 最小二乘法 布谷鳥搜索算法 算法改進(jìn)　出處：《現(xiàn)代制造工程》2017年04期 　論文類型：期刊論文

【摘要】：混合威布爾分布常用于擬合多失效模式的設(shè)備壽命數(shù)據(jù),但由于該分布的形式復(fù)雜且參數(shù)眾多,其參數(shù)估計(jì)較為困難。針對該問題,在對布谷鳥搜索(Cuckoo Search,CS)算法的步長比例和寄主鳥發(fā)現(xiàn)概率改進(jìn)的基礎(chǔ)上,提出基于改進(jìn)CS的混合威布爾分布最優(yōu)化參數(shù)估計(jì)方法。該方法以最小化殘差平方和為目標(biāo),建立參數(shù)估計(jì)優(yōu)化模型,并通過改進(jìn)的CS算法進(jìn)行參數(shù)尋優(yōu)。案例以飛機(jī)擋風(fēng)玻璃壽命數(shù)據(jù)為對象,采用CS算法以及3種改進(jìn)的CS算法分別對兩重兩參數(shù)威布爾分布進(jìn)行2 000次參數(shù)估計(jì),對比分析各算法的尋優(yōu)結(jié)果表明:融合步長比例改進(jìn)和寄主鳥發(fā)現(xiàn)概率改進(jìn)的CS算法的參數(shù)估計(jì)精度較高,估計(jì)結(jié)果較可靠。
[Abstract]:The mixed Weibull distribution is often used to fit the equipment life data of multiple failure modes, but it is difficult to estimate its parameters because of its complex form and numerous parameters. Based on the improvement of step size ratio and host bird discovery probability of Cuckoo SearchCass algorithm, a hybrid Weibull distribution optimization parameter estimation method based on improved CS is proposed. The objective of this method is to minimize the sum of squared residuals. The parameter estimation optimization model is established, and the parameters are optimized by the improved CS algorithm. The case is based on the aircraft windshield life data. The double two-parameter Weibull distribution is estimated by using CS algorithm and three improved CS algorithms for 2,000 times respectively. The comparison and analysis of the optimization results show that the improved CS algorithm with improved step size ratio and host bird discovery probability has higher estimation accuracy and reliable estimation results.
【作者單位】： 軍械工程學(xué)院;
【分類號】：O212.1;TP18

【相似文獻(xiàn)】

相關(guān)期刊論文 前5條

1 宛艷萍,金少華,陸儉國;混合威布爾分布參數(shù)估計(jì)的新算法[J];河北工業(yè)大學(xué)學(xué)報(bào);2002年05期

2 趙繼俊,鄒經(jīng)湘;可靠性實(shí)驗(yàn)數(shù)據(jù)混合威布爾分布的參數(shù)優(yōu)化估計(jì)[J];農(nóng)業(yè)機(jī)械學(xué)報(bào);2000年05期

3 蔣卉;湯銀才;;混合Weibull分布參數(shù)估計(jì)的ECM算法[J];系統(tǒng)科學(xué)與數(shù)學(xué);2010年01期

4 金少華,陸儉國,宛艷萍;關(guān)于參數(shù)估計(jì)的一個(gè)新結(jié)果[J];河北工業(yè)大學(xué)學(xué)報(bào);2000年02期

5 ;[J];;年期

，

本文編號：1530717