本文選題：模糊優(yōu)化 + 滿意度　； 參考：《大連理工大學》2013年博士論文

【摘要】：在國家自然科學基金和國家863高技術(shù)發(fā)展計劃的資助下,本文研究了多學科模糊滿意協(xié)同優(yōu)化方法,主要包括：多學科協(xié)同優(yōu)化(CO)分學科(子系統(tǒng))建模方法、模糊滿意協(xié)同優(yōu)化方法、基于全局共享變量近似子空間的協(xié)同優(yōu)化方法,并將其應(yīng)用于全斷面巖石掘進機刀具布局優(yōu)化設(shè)計中。該問題在理論研究中屬于多學科設(shè)計優(yōu)化方法(MDO)；在機械設(shè)計中屬于復雜布局方案設(shè)計；在人工智能領(lǐng)域中屬于人機結(jié)合問題,具有重要的理論和工程應(yīng)用價值。 提出了基于全局共享變量近似子空間的協(xié)同優(yōu)化方法,對于學科級共享變量的可行域以超切平面法來近似,當系統(tǒng)級共享設(shè)計變量達到了一致性要求,系統(tǒng)級才進行解耦操作,在一定程度上降低學科級的分析次數(shù),具有很好的計算精度和計算穩(wěn)定性,可以有效地解決TBM刀具布局優(yōu)化設(shè)計這類涉及共享變量多,耦合變量相對較少的優(yōu)化設(shè)計問題。 為了能在協(xié)同優(yōu)化中嵌入模糊因素與領(lǐng)域?qū)＜业闹R和經(jīng)驗,本文給出了模糊滿意協(xié)同優(yōu)化模型及算法,在學科級分析模塊中增加了極角模糊規(guī)則和模糊約束最優(yōu)截集水平分析。從TBM刀具布局優(yōu)化設(shè)計問題的應(yīng)用效果來看,有效地解決了刀具布局協(xié)同優(yōu)化設(shè)計中模糊推理、模糊約束的處理問題。 針對TBM刀盤布置設(shè)計中具有布局計算復雜,優(yōu)化器負載重、學科領(lǐng)域內(nèi)容交叉的特點,提出了按MDO的學科分級優(yōu)化的思想,利用協(xié)同優(yōu)化方法,將刀盤最優(yōu)布置模型分成了兩層,共包括1個系統(tǒng)級和4個子系統(tǒng)(學科)級：刀盤幾何學計算子系統(tǒng)、刀盤力學計算子系統(tǒng)、刀盤質(zhì)心計算子系統(tǒng)和破巖量計算子系統(tǒng),各子系統(tǒng)級采用并行優(yōu)化器處理,以降低問題的求解規(guī)模,提高刀具布局尋優(yōu)的效率。 全文的主要工作內(nèi)容包括： (1)討論了并行子空間優(yōu)化、協(xié)同優(yōu)化和兩級集成系統(tǒng)三種多學科優(yōu)化方法的特點和應(yīng)用,并根據(jù)系統(tǒng)工程學原理和TBM盤刀布置問題的特點,提出利用多學科協(xié)同優(yōu)化方法來解決TBM刀具布局優(yōu)化這類復雜問題的思想。 (2)提出了多學科協(xié)同優(yōu)化分學科優(yōu)化策略、建模方法和盤刀布置方案系統(tǒng)級調(diào)整機制。分析了標準協(xié)同優(yōu)化計算困難的原因,在此基礎(chǔ)上提出了一種改良的協(xié)同優(yōu)化方法——基于全局共享變量近似子空間協(xié)同優(yōu)化。并通過實例驗證該方法在求解復雜工程布局問題的可行性和有效性。 (3)提出模糊協(xié)同優(yōu)化模型及算法,主要包括兩類模型——對稱和非對稱模糊協(xié)同優(yōu)化模型,前者將截集水平作為學科級的共享變量來處理,它比較適合于約束條件和優(yōu)化目標具有同等重要程度的情況；后者將截集水平嵌入到學科級中,由學科級分析模塊來處理,比較適合于對約束條件和優(yōu)化目標具有不同的重視度情況。 (4)將本文提出的方法,應(yīng)用到TBM掘進機刀具布局優(yōu)化的工程實例中,具體方法包括：非對稱模糊滿意協(xié)同優(yōu)化模型、學科級分析模塊中極角模糊規(guī)則、模糊約束的最優(yōu)截集水平、學科級模糊推理、極角模糊推理約束條件的構(gòu)建等。 本文以全斷面巖石掘進機刀具布局設(shè)計為應(yīng)用背景,根據(jù)MDO及模糊理論與方法,給出了多學科模糊滿意協(xié)同優(yōu)化方法,并在該方法中應(yīng)用了近似和解耦協(xié)調(diào)機制、模糊優(yōu)化、模糊推理等相關(guān)理論,具有一定的工程實用價值,本文工作期望有助于推動一類復雜工程布局優(yōu)化設(shè)計理論研究的進展。
[Abstract]:With the help of the National Natural Science Foundation and the National 863 High - tech Development Program , this paper studies the multi - disciplinary fuzzy satisfaction collaborative optimization method , which mainly includes the multi - disciplinary collaborative optimization ( CO ) sub - discipline ( subsystem ) modeling method , the fuzzy satisfying cooperative optimization method , the global shared variable approximation subspace and the collaborative optimization method based on the global shared variable approximation subspace , and is applied to the optimization design of the cutter layout of the full - section rock heading machine .
It belongs to the design of complex layout in mechanical design .
In the field of artificial intelligence , it belongs to the combination of man and machine , and has important theory and engineering application value .

In this paper , a collaborative optimization method based on global shared variable approximation subspace is proposed , and the feasible region of the subject - level shared variable is approximated by the hypertangent plane method . When the system - level shared design variable meets the consistency requirement , the system - level is decoupled . It has good calculation accuracy and stability . It can effectively solve the optimization design of TBM cutter layout , which involves many shared variables and relatively few coupling variables .

In order to embed the knowledge and experience of fuzzy factors and domain experts in collaborative optimization , this paper presents a fuzzy satisfying cooperative optimization model and algorithm , which adds polar angle fuzzy rule and fuzzy constraint optimal cutting level analysis in the discipline level analysis module .

According to the characteristics of complex layout calculation , heavy load and cross - cutting in the field of TBM cutter head layout design , the author puts forward the idea of optimization of discipline grading according to MDO . By means of cooperative optimization method , the optimal layout model of cutter head is divided into two layers , including 1 system level and 4 subsystems ( discipline ) level : cutter head geometry calculation subsystem , cutter head mechanics calculation subsystem , cutter head centroid calculation subsystem and rock breaking calculation subsystem , each subsystem level is processed by parallel optimizer to reduce the solution scale of the problem and improve the efficiency of optimization of cutter layout .

The main contents of the full text include :

( 1 ) The characteristics and applications of three multi - disciplinary optimization methods of parallel subspace optimization , cooperative optimization and two - stage integration system are discussed , and the idea of using multi - disciplinary collaborative optimization method to solve the complex problems of TBM cutter layout is proposed based on the characteristics of system engineering principle and TBM disc cutter layout problem .

( 2 ) The optimization strategy , modeling method and system level adjustment mechanism of multi - disciplinary collaborative optimization are put forward . The reasons of difficulty in the calculation of standard collaborative optimization are analyzed . Based on this , an improved cooperative optimization method is proposed , which is based on the global shared variable approximation subspace .

( 3 ) A fuzzy collaborative optimization model and an algorithm are proposed , which mainly include two types of model _ symmetric and asymmetric fuzzy cooperative optimization models . The former deals with the intersection level as the shared variable of the discipline level , which is suitable for the condition that the constraint condition and the optimization goal have the same importance degree .
In the latter , the truncated level is embedded in the discipline level , which is processed by the discipline level analysis module , which is suitable for the condition that the constraint condition and the optimization target have different attention degree .

( 4 ) The method proposed in this paper is applied to the engineering example of TBM boring cutter layout optimization . The concrete method includes : the asymmetric fuzzy satisfying cooperative optimization model , the polar angle fuzzy rule in the discipline level analysis module , the optimal cutting level of the fuzzy constraint , the discipline level fuzzy inference , the construction of the polar angle fuzzy inference constraint condition , and the like .

Based on MDO and fuzzy theory and method , this paper presents a multi - disciplinary fuzzy satisfactory cooperative optimization method based on MDO and fuzzy theory and method .
【學位授予單位】：大連理工大學
【學位級別】：博士
【學位授予年份】：2013
【分類號】：TP18;TH122

【參考文獻】

相關(guān)期刊論文 前10條

1 楊媛媛;黃宏偉;;圍巖分類在TBM滾刀壽命預(yù)測中的應(yīng)用[J];地下空間與工程學報;2005年05期

2 唐曉君,查建中,陸一平;布局問題的復雜性和建模方法[J];北方交通大學學報;2003年01期

3 韓明紅;鄧家y，

本文編號：1950356