摘 要
建筑施工中，工期、質(zhì)量和成本是三個(gè)重要目標(biāo)，這三個(gè)目標(biāo)的好壞對(duì)項(xiàng)目的成功與否具有很大的影響，但是以往對(duì)項(xiàng)目的優(yōu)化大都是考慮工期-成本的優(yōu)化，很少涉及質(zhì)量目標(biāo)，這顯然滿足不了現(xiàn)實(shí)需要。在前人的研究成果中，對(duì)這三個(gè)目標(biāo)進(jìn)行優(yōu)化的模型有兩種：第一種模型是在工期-成本、工期-質(zhì)量、工期-資源為線性關(guān)系的基礎(chǔ)之上構(gòu)建；第二種模型是在工期、成本、質(zhì)量為非線性關(guān)系的基礎(chǔ)之上構(gòu)建。
本文分別就這兩種不同的關(guān)系提出了兩種綜合優(yōu)化模型，第一種模型是根據(jù)質(zhì)量、成本與時(shí)間的關(guān)系，建立成本和質(zhì)量的目標(biāo)函數(shù)，然后求出各工序的可能完成時(shí)間所對(duì)應(yīng)的成本和質(zhì)量。為了應(yīng)用遺傳算法進(jìn)行求解我們把各個(gè)工序的可能完成時(shí)間以及其所對(duì)應(yīng)的質(zhì)量和所需費(fèi)用作為一個(gè)模式，每個(gè)工序的可能完成時(shí)間即對(duì)應(yīng)著其模式的數(shù)量，這樣每個(gè)模式都對(duì)應(yīng)不同的施工時(shí)間、施工質(zhì)量和施工成本，最后由每個(gè)工序的模式組成染色體，通過(guò)遺傳算法進(jìn)行求解；在第二種模型中，完成各個(gè)工序的時(shí)間、成本和質(zhì)量是沒(méi)有聯(lián)系的，每個(gè)工序的不同完成時(shí)間對(duì)應(yīng)相應(yīng)的成本和質(zhì)量，其染色體的組成同第一種模型，然后用遺傳算法進(jìn)行求解，運(yùn)行之后可以得到多組解，決策者可以根據(jù)自己的偏好進(jìn)行選擇，最后對(duì)這兩種模型進(jìn)行比較，以確定出彼此的優(yōu)缺點(diǎn)。
為了避免資源在單位時(shí)間內(nèi)投入量過(guò)大，在三大目標(biāo)綜合優(yōu)化的基礎(chǔ)上加入了資源均衡操作，提出了一個(gè)兩階段優(yōu)化模型，第一階段是對(duì)工期、成本和質(zhì)量進(jìn)行綜合優(yōu)化；第二階段是從第一階段對(duì)所得到的非劣解中，由決策者選擇一個(gè)或多個(gè)滿意的解輸入到本階段進(jìn)行資源均衡優(yōu)化，把各個(gè)工序的開(kāi)始時(shí)間作為模式構(gòu)成染色體，進(jìn)行遺傳操作，最后通過(guò)一個(gè)工程案例證明了兩階段模型的可行性與優(yōu)越性。
關(guān)鍵詞：質(zhì)量；資源；綜合優(yōu)化；遺傳算法；模式

Abstract
Duration, cost and quality is the three main objectives of the construction. The performance of these three objectives can directly affect the entire target of the project .The best solution for a project is to achieve high quality with low cost in a proper time limit. But the previous model which could only resolve the time-cost trade-off problems that cannot satisfy decision makers’ requirement. Because of these three goals are conflicting, and if one of the goal is improvement, other goals will be weakened inevitably. So therefore we can only enable the overall goal of project to achieve superiorly , that is to say under the premise of the total target is satisfied, to ensure the quality is relatively high, relatively short duration and the cost is relatively low.
As a result of the limitation of the method, and most of the optimal method is about two goals of the three objectives, so we should find a method which is contains three goals. In predecessor’s research results, there are to types of synthesis optimization models: and one of the models is established on the base of the relation of time-cost, time-quality is linear relationship; and the other model is established on the base of the relation of time-cost, time-quality is unlinear relationship. And this paper proposed two synthesis optimization models which should be soluted by Genetic algorithm is established on the base of these two types of relation. With this method you can obtain some solutions, and the policy-maker can choose one of it which is satisfied himself. And these two methods will be compared with each other, and the way of how to choose will be pointed out.
Simultaneously in order to avoid the inputs of resource is overised in unit of time ,we proposed a two stage optimization model ,the first stage established a synthesis optimization models which is based on the relationship of the time, the quality and the cost is the misalignment, and it uses chromosomes which building up by mode to carry on the heredity operation; The second stage is to choose a non-poor solution which obtains from the Decision-maker from the first stage inputs to this stage to carries on the resource leveling optimization, it take the time in resources variance minimum as the objective function, takes the gene value which is non-critical process's beginning time to make up of the chromosome to carry on the heredity operation, An application example is analyzed to illustrate the use of the model and demonstrate its capabilities in generating and visualizing optimal tradeoffs among construction time, cost and quality, and resource.

Key words：Comprehensive optimization； Schema； GA；Quality; Resource
目錄
摘要 I
Abstract II
第一章 緒 論 1
1.1 研究的背景及意義 1
1.1.1 研究背景 1
1.1.2 研究的意義 2
1.2 國(guó)內(nèi)外研究現(xiàn)狀與發(fā)展趨勢(shì) 2
1.2.1多目標(biāo)優(yōu)化模型的研究現(xiàn)狀 2
1.2.2 遺傳算法的研究現(xiàn)狀 7
1.2.3對(duì)已有研究成果評(píng)述 7
1.3 課題研究?jī)?nèi)容、研究方法及技術(shù)路線 8
1.3.1研究?jī)?nèi)容 8
1.3.2研究方法 9
1.4 論文的創(chuàng)新點(diǎn) 10
第二章 遺傳算法的原理及實(shí)現(xiàn)方法 11
2.1 遺傳算法的研究現(xiàn)狀 11
2.2遺傳算法的基本要素 13
2.3遺傳算法的特點(diǎn) 15
2.3.1遺傳算法的優(yōu)點(diǎn)： 15
2.3.2遺傳算法的缺點(diǎn) 16
2.4 遺傳算法的操作流程 16
2.5 本章小結(jié) 17
第三章 工期-成本-質(zhì)量為線性關(guān)系的綜合優(yōu)化 18
3.1 工期、質(zhì)量、成本之間的關(guān)系 18
3.1.1 工期-成本之間的關(guān)系 18
3.1.2 質(zhì)量-成本之間的關(guān)系 19
3.1.3 三大目標(biāo)之間的相互關(guān)系 20
3.2 工期、成本、質(zhì)量為線性關(guān)系的綜合優(yōu)化模..