【摘要】：響應(yīng)曲面方法(RSM)是現(xiàn)代統(tǒng)計(jì)與數(shù)學(xué)的集成方法,被廣泛應(yīng)用于響應(yīng)變量與自變量之間函數(shù)關(guān)系研究中。通過響應(yīng)曲面設(shè)計(jì)、建模和優(yōu)化完成整個(gè)響應(yīng)曲面過程。在響應(yīng)曲面建模中,通常要求隨機(jī)誤差服從正態(tài)分布相互獨(dú)立且具有相同的方差;輸入變量相互獨(dú)立,沒有測(cè)量誤差。在實(shí)際生產(chǎn)過程中,這些假設(shè)條件只能近似地被滿足,并且自變量中很可能會(huì)出現(xiàn)測(cè)量誤差,響應(yīng)數(shù)據(jù)中也可能存在異常點(diǎn)情況。針對(duì)上述兩個(gè)問題,本文主要研究在輸入變量存在測(cè)量誤差時(shí),響應(yīng)曲面設(shè)計(jì)的穩(wěn)健性以及在異常點(diǎn)存在時(shí),響應(yīng)曲面模型的穩(wěn)健性。 本文介紹了響應(yīng)曲面方法的一些基本理論和傳統(tǒng)的模型估計(jì)方法、響應(yīng)曲面設(shè)計(jì)的評(píng)價(jià)準(zhǔn)則、穩(wěn)健估計(jì)方法和誤差變量模型。在此基礎(chǔ)了比較分析了不同的響應(yīng)曲面設(shè)計(jì)和不同估計(jì)方法的穩(wěn)健性,具體從以下幾個(gè)方面進(jìn)行研究。 首先,探討了不同的響應(yīng)曲面設(shè)計(jì)對(duì)響應(yīng)數(shù)據(jù)中異常點(diǎn)的穩(wěn)健性問題。主要包括中心復(fù)合設(shè)計(jì)——中心復(fù)合序貫設(shè)計(jì)、中心復(fù)合有界設(shè)計(jì)和Box-Behnken設(shè)計(jì)�；谶@些設(shè)計(jì)的比較分析結(jié)果表明,中心點(diǎn)位置異常點(diǎn)極大地增加了模型的波動(dòng)性,降低了模型的精度,軸線點(diǎn)和立方體點(diǎn)位置異常點(diǎn)影響模型二次項(xiàng)或交叉項(xiàng)的估計(jì),增加了模型優(yōu)化和預(yù)測(cè)的難度。 其次,提出了兩種響應(yīng)曲面設(shè)計(jì)預(yù)測(cè)方差的評(píng)價(jià)指標(biāo),研究分析了誤差變量模型的單位預(yù)測(cè)方差和傳統(tǒng)的三因子響應(yīng)曲面設(shè)計(jì)預(yù)測(cè)性能。仿真分析表明,對(duì)于測(cè)量誤差的方差比較小時(shí),中心復(fù)合表面設(shè)計(jì)的穩(wěn)健性是最優(yōu)的;當(dāng)方差較大時(shí),Box-Behnken設(shè)計(jì)的預(yù)測(cè)性更好一些。 再次,對(duì)存在測(cè)量誤差的二階球形響應(yīng)曲面設(shè)計(jì)優(yōu)化準(zhǔn)則和預(yù)測(cè)方差性質(zhì)進(jìn)行比較和評(píng)估。為了直觀展示不同設(shè)計(jì)對(duì)測(cè)量誤差的穩(wěn)健性,我們提出單位預(yù)測(cè)方差最大值與最小值圖和設(shè)計(jì)空間比率圖,計(jì)算了不同設(shè)計(jì)的G-效率值。得出了不同因子水平下,穩(wěn)健的響應(yīng)曲面設(shè)計(jì)方案。 最后,基于響應(yīng)數(shù)據(jù)中的異常點(diǎn),探究了二階響應(yīng)曲面模型的穩(wěn)健估計(jì)方法,研究分析了異常點(diǎn)存在或非正態(tài)時(shí)不同估計(jì)方法的穩(wěn)健性。通過仿真和實(shí)例得出比較結(jié)果。
[Abstract]:Response surface method (RSM) is an integrated method of modern statistics and mathematics, which is widely used in the study of the functional relationship between response variables and independent variables. The whole process of response surface is completed by designing, modeling and optimizing the response surface. In response surface modeling, random errors are usually required to be independent from normal distribution and have the same variance, and input variables are independent without measurement errors. In the actual production process, these assumptions can only be approximately satisfied, and there may be measurement errors in the independent variables, and there may be outliers in the response data. In view of the above two problems, the robustness of response surface design and the robustness of response surface model when there are measurement errors in input variables are studied in this paper. This paper introduces some basic theories of response surface method and traditional model estimation method, evaluation criterion of response surface design, robust estimation method and error variable model. Based on this, the robustness of different response surface design and different estimation methods are compared and analyzed. Firstly, the robustness of different response surface designs to outliers in response data is discussed. It mainly includes central composite design, central composite sequential design, center composite bounded design and Box-Behnken design. The results of comparative analysis based on these designs show that the center point outliers greatly increase the volatility of the model and reduce the accuracy of the model. The location anomaly points of the axis and cube points affect the estimation of the quadratic or cross terms of the model. It increases the difficulty of model optimization and prediction. Secondly, two evaluation indexes of prediction variance of response surface design are proposed, and the unit prediction variance of error variable model and the prediction performance of traditional three-factor response surface design are studied and analyzed. Simulation results show that the robustness of the central composite surface design is optimal when the variance ratio of the measurement error is small, and the predictability of the Box-Behnken design is better when the variance is large. Thirdly, the optimization criteria and the properties of predictive variance for the design of second-order spherical response surfaces with measurement errors are compared and evaluated. In order to show the robustness of different designs to the measurement error, we propose the maximum and minimum values of unit predictive variance and the design space ratio diagram, and calculate the G- efficiency values of different designs. A robust response surface design scheme with different factor levels is obtained. Finally, based on the outliers in the response data, the robust estimation methods of the second-order response surface model are discussed, and the robustness of the different estimation methods for the existence or non-normal state of the outliers is analyzed. The comparison results are obtained by simulation and examples.
【學(xué)位授予單位】：天津大學(xué)
【學(xué)位級(jí)別】：博士
【學(xué)位授予年份】：2011
【分類號(hào)】：O211.3;C81

【參考文獻(xiàn)】

相關(guān)期刊論文 前10條

1 張志紅;何楨;;工業(yè)工程(第六講) 實(shí)驗(yàn)設(shè)計(jì)與質(zhì)量改進(jìn)[J];工程機(jī)械;2006年02期

2 何楨,張于軒;多響應(yīng)試驗(yàn)設(shè)計(jì)的優(yōu)化方法研究[J];工業(yè)工程;2003年04期

3 何楨;崔慶安;;基于支持向量機(jī)的小樣本響應(yīng)曲面法研究[J];工業(yè)工程;2006年05期

4 何楨;馬彥輝;趙有;;非正態(tài)響應(yīng)的部分因子試驗(yàn)設(shè)計(jì)[J];工業(yè)工程;2008年02期

5 何楨;張凱;梁昭磊;;關(guān)聯(lián)因子裂區(qū)的響應(yīng)曲面設(shè)計(jì)[J];工業(yè)工程;2009年03期

6 繳錫云;雷志棟;彭世彰;;建立作物水分生產(chǎn)函數(shù)的穩(wěn)健回歸方法[J];灌溉排水學(xué)報(bào);2006年03期

7 朱鵬飛;何楨;;混合模擬退火的多響應(yīng)參數(shù)優(yōu)化[J];工業(yè)工程與管理;2010年04期

8 麻金繼;李素文;;基于穩(wěn)健回歸M估計(jì)的差分吸收光譜反演方法(英文)[J];光子學(xué)報(bào);2009年08期

9 崔慶安;何楨;崔楠;;基于SVM的RSM模型擬合方法研究[J];管理科學(xué)學(xué)報(bào);2008年01期

10 周劍忠;李瑩;單成俊;王英;張麗霞;劉小莉;;響應(yīng)曲面法優(yōu)化植物乳桿菌蔬菜汁培養(yǎng)基的研究[J];江蘇農(nóng)業(yè)科學(xué);2008年05期

相關(guān)博士學(xué)位論文 前4條

1 張志紅;基于并行質(zhì)量工程的6σ公差設(shè)計(jì)方法的研究[D];天津大學(xué);2005年

2 崔慶安;基于支持向量回歸機(jī)的復(fù)雜過程響應(yīng)曲面法研究[D];天津大學(xué);2007年

3 宗志宇;基于RSM的多響應(yīng)穩(wěn)健性設(shè)計(jì)方法的研究[D];天津大學(xué);2007年

4 王晶;基于響應(yīng)曲面法的多響應(yīng)穩(wěn)健性參數(shù)優(yōu)化方法研究[D];天津大學(xué);2009年

，

本文編號(hào)：2389463