本文關(guān)鍵詞：生態(tài)金字塔粒子群優(yōu)化算法及其在蛋殼薄膜膠原蛋白提取中的應(yīng)用　出處：《太原理工大學(xué)》2017年碩士論文　論文類型：學(xué)位論文

  更多相關(guān)文章： 粒子群算法 生態(tài)金字塔系統(tǒng) 膠原蛋白 響應(yīng)面方法 試驗設(shè)計

【摘要】：粒子群優(yōu)化算法(Particle Swarm Optimization,PSO)最早由Eberhart和Kennedy在1995年提出,是一種在解決多種優(yōu)化問題中得到廣泛應(yīng)用和發(fā)展的群智能算法。在粒子群優(yōu)化算法中,優(yōu)化問題的可能解被視為鳥群的食物目標(biāo),群體中的粒子通過信息共享和互助機制,引導(dǎo)整個群體朝著可能解的位置運動,在該過程中逐漸找到更好的全局最優(yōu)解。PSO算法因具備結(jié)構(gòu)簡單、收斂迅速和易于編程實現(xiàn)的特性,獲得研究者們的廣泛關(guān)注,短短二十幾年里迅速發(fā)展成進化算法的一個重要分支,在多個領(lǐng)域得到廣泛應(yīng)用。雖然學(xué)者們從不同方面對粒子群優(yōu)化算法進行了改進,提出多種改進算法,并已經(jīng)取得了一定的成果。但在解決高維復(fù)雜函數(shù)時仍然存在問題,如容易陷入局部最優(yōu)、過早收斂以及低精度等,因此,在求解此類問題時,粒子群優(yōu)化算法的性能仍有待改進和提高。針對上述問題,本文對粒子群算法的結(jié)構(gòu)特點和搜索過程進行了深入分析和探討,為克服粒子群優(yōu)化算法處理高維復(fù)雜函數(shù)容易陷入局部最優(yōu)和過早收斂的問題,提出了生態(tài)金字塔粒子群優(yōu)化算法(EP-PSO),并將EP-PSO算法應(yīng)用于蛋殼薄膜膠原蛋白的提取工藝中,優(yōu)化響應(yīng)面回歸模型,確定最優(yōu)的實驗因素和水平。主要研究內(nèi)容如下:1.為克服粒子群優(yōu)化算法處理高維復(fù)雜函數(shù)容易陷入局部最優(yōu)和早熟收斂的問題,本文提出生態(tài)金字塔粒子群優(yōu)化算法(EP-PSO)。該算法引入生態(tài)金字塔系統(tǒng),使粒子在搜索空間分等級、分子群尋優(yōu),有效提高了群體多樣性;為增強算法的全局搜索能力,對處于停滯狀態(tài)的個體極值和全局極值進行動態(tài)變異,達到擴大種群潛在搜索空間的效果。并選取了9種經(jīng)典粒子群改進算法和15個標(biāo)準(zhǔn)測試函數(shù)對生態(tài)金字塔粒子群算法進行了對比試驗分析,結(jié)果表明EP-PSO有著良好的尋優(yōu)性能,能夠得到較高精度解,具有較高的效率和可信度。2.隨著生物科學(xué)發(fā)展進程的不斷推進,膠原蛋白越來越受到大家的關(guān)注,本文將EP-PSO應(yīng)用于蛋殼薄膜膠原蛋白提取工藝的優(yōu)化中,通過仿真試驗驗證了EP-PSO的有效性,證明了氫氧化鈉濃度、堿處理時間、酶濃度和水解時間四因素為影響從蛋殼薄膜中提取膠原蛋白的主要因素,得到四因素的適宜取值范圍分別為:氫氧化鈉濃度(A):0.7mol/L~0.8mol/L、堿處理時間(B):12h~20h、酶濃度(C):25U/mg~60U/mg和水解時間(D):36h~48h。
[Abstract]:Particle swarm optimization algorithm (Particle Swarm Optimization, PSO) first by Eberhart and Kennedy in 1995, is a kind of swarm intelligence algorithm to solve the application and development of a variety of optimization problems. In the particle swarm optimization algorithm, the optimization solution of the problem is regarded as the bird food, in the group the particle through information sharing and mutual assistance mechanism, guide the population toward the position of moving the possible solutions, and gradually find a better global optimal in the process solution for.PSO algorithm has the advantages of simple structure, fast convergence and easy programming features, received wide attention of researchers, only more than 20 years of rapid development into a an important branch of the evolutionary algorithm, is widely used in many fields. Although scholars from different aspects of the particle swarm optimization algorithm, proposed improved algorithm, and has achieved a The results. But there are still problems in solving high dimension complex functions, such as easy to fall into local optimum, premature convergence and low precision, therefore, in solving these problems, the performance of particle swarm optimization still needs to be improved. Aiming at the above problems, the structure characteristics and the search process of particle swarm algorithm the in-depth analysis and discussion, in order to overcome the particle swarm optimization algorithm for complex functions with high dimension is easy to fall into local optimum and premature convergence problem, put forward the ecological Pyramid particle swarm optimization algorithm (EP-PSO), and the EP-PSO algorithm is applied to the extraction process of eggshell thin film of collagen, optimization of response surface regression model, determine the experimental factors and the optimal level. The main contents are as follows: 1. in order to overcome the particle swarm optimization algorithm for complex functions with high dimension is easy to fall into local optimum and premature convergence problem, is proposed in this paper. Pyramid ecological particle swarm optimization (EP-PSO) algorithm is introduced. The ecological system of Pyramid, the particles in the search space level, molecular swarm optimization, to improve the population diversity; to enhance the global search capability of the algorithm, the dynamic variation of individual extremum and global extremum in a state of stagnation, to expand the population of potential search the effects of space. And select 9 kinds of classical particle swarm algorithm and 15 standard test functions are analyzed in comparison to the ecological Pyramid particle swarm algorithm, the results show that EP-PSO has a good optimization, we can get high precision solution with high efficiency and reliability of.2. with the progress of the development of biological science collagen, has attracted more and more attention, this article will optimize the application of EP-PSO in the eggshell film collagen extraction process, through simulation and experimental verification of the EP-P The effectiveness of SO, proved that the concentration of sodium hydroxide, alkali treatment time, enzyme concentration and hydrolysis time four factors as the main factors affecting the extraction of collagen from eggshell membrane, get the appropriate range of the four factors were: the concentration of sodium hydroxide (A): 0.7mol/L~0.8mol/L, alkali treatment time (B): 12h~20h (C, enzyme concentration: 25U/mg~60U/mg) and hydrolysis time (D): 36h~48h.

【學(xué)位授予單位】：太原理工大學(xué)
【學(xué)位級別】：碩士
【學(xué)位授予年份】：2017
【分類號】：TQ936.2;TP18

【參考文獻】

相關(guān)期刊論文 前8條

1 高云龍;閆鵬;;基于多種群粒子群算法和布谷鳥搜索的聯(lián)合尋優(yōu)算法[J];控制與決策;2016年04期

2 王瀝浩;王文慧;郭詠昕;楊晶;;膠原蛋白功能概述[J];黑龍江農(nóng)業(yè)科學(xué);2014年03期

3 趙遠東;方正華;;帶有權(quán)重函數(shù)學(xué)習(xí)因子的粒子群算法[J];計算機應(yīng)用;2013年08期

4 潘峰;周倩;李位星;高琪;;標(biāo)準(zhǔn)粒子群優(yōu)化算法的馬爾科夫鏈分析[J];自動化學(xué)報;2013年04期

5 俞藝萍;陸利霞;熊曉輝;;魚皮膠原蛋白提取研究進展[J];食品研究與開發(fā);2010年12期

6 鄒德旋;王高楠;高立群;吳建華;;改進的粒子群優(yōu)化算法在可靠性問題中的應(yīng)用[J];東北大學(xué)學(xué)報(自然科學(xué)版);2010年09期

7 張俊杰;段蕊;許曉靜;李晶;鐘曉紅;許兵;;魷魚魚皮酶溶性膠原蛋白部分生化性質(zhì)研究[J];食品與發(fā)酵工業(yè);2009年12期

8 胡旺;李志蜀;;一種更簡化而高效的粒子群優(yōu)化算法[J];軟件學(xué)報;2007年04期

相關(guān)博士學(xué)位論文 前1條

1 陳洋;基于高斯動態(tài)PSO算法的艦船電網(wǎng)故障重構(gòu)方法研究[D];大連海事大學(xué);2015年

相關(guān)碩士學(xué)位論文 前5條

1 馬迪;粒子群響應(yīng)面建模法在ASPEN多因素優(yōu)化中的應(yīng)用[D];太原理工大學(xué);2016年

2 劉巋;響應(yīng)面法優(yōu)化超聲輔助提取銀杏葉中總黃酮及其抗氧化研究[D];蘭州大學(xué);2016年

3 魯寧;雞胸軟骨膠原蛋白制備工藝研究[D];安徽農(nóng)業(yè)大學(xué);2013年

4 張盈盈;基于粒子群算法的非概率可靠性分析及其在邊坡工程中的應(yīng)用[D];中南大學(xué);2011年

5 胡雅琴;響應(yīng)曲面二階設(shè)計方法比較研究[D];天津大學(xué);2005年

，

本文編號：1416969