【摘要】：基因表達(dá)式編程算法(Gene Expression Programming,GEP)是一種新型的處理高維的、不確定性因素的智能進(jìn)化算法,它能夠挖掘出隱藏在數(shù)據(jù)中的知識(shí),如規(guī)則、模型等,并且不需要任何的先驗(yàn)知識(shí)。該算法以獨(dú)特的編碼方式、優(yōu)秀的數(shù)據(jù)挖掘能力和高度非線性系統(tǒng)的處理能力吸引了許多國(guó)內(nèi)外研究者的關(guān)注,并廣泛應(yīng)用于眾多的實(shí)際領(lǐng)域中。本文主要的研究工作是對(duì)標(biāo)準(zhǔn)GEP算法進(jìn)行改進(jìn),并將其應(yīng)用到知識(shí)發(fā)現(xiàn)中的兩大問(wèn)題中,即麥蚜種群建模和建筑工程造價(jià)預(yù)測(cè),可為麥蚜種群發(fā)生量的預(yù)測(cè)和建筑項(xiàng)目的可行性研究以及合理的設(shè)計(jì)方案提供依據(jù)。本研究的具體工作有以下幾點(diǎn):(1)在閱讀大量文獻(xiàn)和建模預(yù)測(cè)內(nèi)容的基礎(chǔ)上,本文對(duì)應(yīng)用領(lǐng)域的重點(diǎn)知識(shí)進(jìn)行總結(jié)概括,介紹了麥蚜種群建模的過(guò)程和現(xiàn)狀以及建筑工程造價(jià)預(yù)測(cè)特征量的提取與分類;對(duì)基因表達(dá)式編程算法的原理和編碼方式進(jìn)行了概述,詳細(xì)介紹了算法的流程和基本操作。(2)依據(jù)人工干預(yù)思想,本研究提出了一種由人工干預(yù)系統(tǒng)和自然進(jìn)化系統(tǒng)組成的雙系統(tǒng)協(xié)同進(jìn)化的基因表達(dá)式編程算法(DSCE-GEP)。人工干預(yù)系統(tǒng)包括個(gè)體干預(yù)和種群干預(yù),個(gè)體干預(yù)即采用優(yōu)質(zhì)基因庫(kù)對(duì)種群中的個(gè)體進(jìn)行增優(yōu)和去劣操作,旨在提高個(gè)體的質(zhì)量;種群干預(yù)則是利用信息熵通過(guò)引入隨機(jī)個(gè)體和鏡像個(gè)體來(lái)提高種群多樣性。(3)針對(duì)提出的改進(jìn)算法,本研究與類似算法進(jìn)行了對(duì)比仿真實(shí)驗(yàn),分析、驗(yàn)證了DSCE-GEP算法的有效性和先進(jìn)性。同時(shí),本文將DSCE-GEP算法應(yīng)用于農(nóng)業(yè)和建筑業(yè)中,對(duì)中國(guó)農(nóng)業(yè)科學(xué)院提供的麥蚜種群數(shù)據(jù)進(jìn)行建模,并對(duì)文獻(xiàn)中列舉的建筑工程項(xiàng)目數(shù)據(jù)進(jìn)行建模預(yù)測(cè)。實(shí)驗(yàn)結(jié)果表明,本研究構(gòu)建的基于基因表達(dá)式編程算法的麥蚜種群模型和建筑工程造價(jià)預(yù)測(cè)模型,建模效果優(yōu)越,預(yù)測(cè)精度較高。
[Abstract]:Gene expression programming algorithm (Gene Expression Programming,GEP) is a new kind of intelligent evolutionary algorithm dealing with high dimensional and uncertain factors. It can mine hidden knowledge in data, such as rules, models, etc. And no prior knowledge is required. This algorithm has attracted the attention of many researchers at home and abroad with its unique coding method, excellent data mining ability and processing ability of highly nonlinear systems, and has been widely used in many practical fields. The main work of this paper is to improve the standard GEP algorithm and apply it to the two major problems of knowledge discovery, that is, wheat aphid population modeling and construction engineering cost prediction. It can provide basis for prediction of population occurrence of wheat aphid, feasibility study of construction project and reasonable design scheme. The specific work of this study is as follows: (1) on the basis of reading a lot of literature and modeling and forecasting content, this paper summarizes the key knowledge of the application field. This paper introduces the process and present situation of the population modeling of wheat aphid, and the extraction and classification of the characteristic quantity of construction engineering cost prediction. The principle and coding method of gene expression programming algorithm are summarized, and the flow and basic operation of the algorithm are introduced in detail. (2) according to the idea of human intervention, In this paper, a gene expression programming algorithm (DSCE-GEP) is proposed, which consists of artificial intervention system and natural evolution system. The artificial intervention system includes individual intervention and population intervention. Individual intervention is to improve the quality of individuals by using high quality gene pool to improve the quality of individuals. Population intervention is to use information entropy to improve population diversity by introducing random individuals and mirrored individuals. (3) aiming at the proposed improved algorithm, this paper makes a comparative simulation experiment with similar algorithms. The validity and advancement of DSCE-GEP algorithm are verified. At the same time, the DSCE-GEP algorithm is applied to agriculture and construction, to model the wheat aphid population data provided by the Chinese Academy of Agricultural Sciences, and to model and predict the construction project data listed in the literature. The experimental results show that the population model of wheat aphid based on genetic expression programming algorithm and the cost prediction model of construction engineering are superior to each other and the precision of prediction is high.
【學(xué)位授予單位】：西安建筑科技大學(xué)
【學(xué)位級(jí)別】：碩士
【學(xué)位授予年份】：2016
【分類號(hào)】：TP18
，

本文編號(hào)：2382194