本文選題：可測(cè)試性設(shè)計(jì)　切入點(diǎn)：測(cè)試點(diǎn)優(yōu)化選擇　出處：《光學(xué)精密工程》2017年05期

【摘要】：測(cè)試點(diǎn)優(yōu)化選擇是復(fù)雜裝備測(cè)試性設(shè)計(jì)的重要環(huán)節(jié),本文提出一種用于解決測(cè)試點(diǎn)優(yōu)化選擇問(wèn)題的離散螢火蟲(chóng)算法(DFA)。首先建立了測(cè)試點(diǎn)優(yōu)化選擇問(wèn)題的數(shù)學(xué)模型,接著對(duì)傳統(tǒng)的螢火蟲(chóng)算法(FA)進(jìn)行了離散化改進(jìn),給出了離散化螢火蟲(chóng)算法的實(shí)施步驟,并分析了不同的吸引度函數(shù)和二值化函數(shù)(sigmoid和tanh函數(shù))對(duì)算法結(jié)果的影響。最后針對(duì)5個(gè)不同規(guī)模的實(shí)際系統(tǒng)驗(yàn)證了離散螢火蟲(chóng)算法的有效性,并與粒子群算法(PSO)和遺傳算法(GA)等傳統(tǒng)的元啟發(fā)式搜索算法的計(jì)算性能進(jìn)行了比較分析。結(jié)果顯示:在滿足系統(tǒng)要求的故障檢測(cè)率和故障隔離率的前提下,利用本文提出的離散螢火蟲(chóng)算法得到的5個(gè)系統(tǒng)測(cè)試代價(jià)最優(yōu)值分別比PSO算法和GA算法平均降低了10.1%和14.6%。實(shí)驗(yàn)結(jié)果表明:離散螢火蟲(chóng)算法能快速收斂到更高質(zhì)量的全局最優(yōu)解,避免過(guò)早收斂而陷入局部最優(yōu)值,對(duì)于解決大型復(fù)雜裝備的測(cè)試點(diǎn)優(yōu)化選擇問(wèn)題具有很好的應(yīng)用前景。
[Abstract]:The optimal selection of test points is an important link in the testability design of complex equipment. In this paper, a discrete firefly algorithm is proposed to solve the problem of optimal selection of test points.Firstly, the mathematical model of optimal selection of test points is established, and then the discretization of the traditional firefly algorithm (FAA) is improved, and the implementation steps of the discrete firefly algorithm are given.The effects of different attraction functions and binarization functions (sigmoid and tanh functions) on the results of the algorithm are analyzed.Finally, the effectiveness of the discrete firefly algorithm is verified for five different scale systems, and the computational performance of the traditional meta-heuristic search algorithm, such as particle swarm optimization (PSO) and genetic algorithm (GA), is compared and analyzed.The results show that under the premise of fault detection rate and fault isolation rate which meet the requirements of the system, the optimal test cost of the five systems obtained by using the discrete firefly algorithm proposed in this paper is 10.1% and 14.6% lower than that of the PSO algorithm and the GA algorithm, respectively.The experimental results show that the discrete firefly algorithm can quickly converge to a higher quality global optimal solution, avoid premature convergence and fall into a local optimal value, and has a good application prospect for solving the problem of test point selection for large and complex equipment.
【作者單位】： 中國(guó)科學(xué)院長(zhǎng)春光學(xué)精密機(jī)械與物理研究所;
【基金】：國(guó)家重點(diǎn)實(shí)驗(yàn)室研究基金資助項(xiàng)目(No.SKLLIM0902-01)
【分類號(hào)】：TP18
，

本文編號(hào)：1724321