丁根宏 曹文秀
摘要:水庫防洪優(yōu)化調(diào)度模型一般屬于高維多峰極值問題,通常采用智能優(yōu)化算法加以求解。粒子群算法由于其簡單易行被廣泛應(yīng)用于水庫優(yōu)化調(diào)度中,但是該算法存在局部搜索能力不足、早熟收斂、全局收斂性差等問題。針對這些問題,通過引入Logistic方程和變異算子來提高種群的多樣性,采用收斂因子來提高算法的收斂速度,并將改進(jìn)的粒子群算法應(yīng)用到東圳水庫與木蘭溪流域的防洪優(yōu)化調(diào)度中,求得關(guān)鍵處河道的最高水位為6.35 m,最大流量為959.2 m3/s。這一結(jié)果與現(xiàn)行規(guī)則下的運(yùn)行結(jié)果(最高水位6.93 m,最大流量1 139.5 m3/s)和常規(guī)粒子群算法計(jì)算結(jié)果(最高水位6.51 m,最大流量1 066.3 m3/s)相比,有了很大的改善。
關(guān)鍵詞:防洪調(diào)度;智能優(yōu)化;粒子群算法;混沌思想;變異策略;收斂因子
中圖分類號:TV697 文獻(xiàn)標(biāo)識碼:A 文章編號:1672-1683(2014)01-0118-04
我國是世界上洪澇災(zāi)害發(fā)生頻率最高、受災(zāi)最重的少數(shù)國家之一[1],水庫防洪問題已成為學(xué)術(shù)界普遍關(guān)注的問題[2-3]。目前在水庫調(diào)度過程中,除了采用具有固定調(diào)度規(guī)則的常規(guī)調(diào)度外,普遍采用最優(yōu)化方法和現(xiàn)代計(jì)算技術(shù)來求解以水庫為中心的滿足一定約束條件的數(shù)學(xué)模型,實(shí)現(xiàn)水庫的優(yōu)化調(diào)度[4-6]。相較于遺傳算法和蟻群算法,粒子群算法的簡單易實(shí)現(xiàn)、收斂速度快等特點(diǎn)更適合于水庫優(yōu)化調(diào)度模型的求解[7-8]。本文在前人研究的基礎(chǔ)之上,對粒子群算法及其在水庫防洪中的應(yīng)用作了進(jìn)一步的探索和研究,旨在為解決水庫優(yōu)化調(diào)度問題提供一些新理論和新方法及相應(yīng)的實(shí)證分析。
1 粒子群算法的改進(jìn)
粒子群算法(PSO)是一種進(jìn)化計(jì)算技術(shù)(evolutionary computation),最早是在1995年由美國社會心理學(xué)家James Kennedy博士和電氣工程師Russell Ebethart博士受人工生命和演化計(jì)算理論的研究結(jié)果的啟發(fā)提出的[9-10]。PSO模擬鳥群隨機(jī)搜索食物的行為,將鳥群的搜索區(qū)域?qū)?yīng)于設(shè)計(jì)[HJ]變量的變化范圍,食物對應(yīng)于適應(yīng)度函數(shù)的最優(yōu)解[11-12]。本文主要從提高種群的多樣性和提高收斂速度兩個(gè)方面對粒子群算法進(jìn)行改進(jìn)。
3結(jié)語
本文提出的改進(jìn)粒子群算法運(yùn)用到東圳水庫及木蘭溪流域防洪優(yōu)化調(diào)度中,得到的水庫下泄過程使A10處最高水位為6.35 m,最大流量為959.2 m3/s,該結(jié)果明顯優(yōu)于采用現(xiàn)行水庫調(diào)度方案所得到的A10處最高水位6.93 m以及最大流量1 139.5 m3/s,也優(yōu)于采用常規(guī)粒子群算法調(diào)度方案所得到的A10處最高水位6.51 m以及最大流量1 066.3 m3/s,表明改進(jìn)粒子群算法在水庫防洪調(diào)度中有很好的應(yīng)用,充分顯示了改進(jìn)粒子群算法的優(yōu)良性能。
參考文獻(xiàn)(References):
[1] 毛德華,邱永紅,謝石,等.洪災(zāi)風(fēng)險(xiǎn)分析的國內(nèi)外研究現(xiàn)狀及展望(Ⅲ)—研究展望[J].自然災(zāi)害學(xué)報(bào),2012,(5):8-15.(MAO De-hua,QIU Yong-hong,XIE Shi,et al.Review and Prospect of Research on Flooding Risk Analysis at Home and Abroad (Ⅲ):Research Prospect[J].Journal of Natural Disasters,2012,(5):8-15.(in Chinese))
[2] 李立,朱毅.工程與非工程措施并舉構(gòu)建現(xiàn)代防洪體系[J].湖南水利水電,2004,(3):36-37.(LI Li,ZHU Yi.The Construction of Modern Flood Control System in Engineering and Non-engineering Measures[J].Hunan Hydro & Power,2004,(3):36-37.(in Chinese))
[3] 鄒進(jìn).水庫長期優(yōu)化調(diào)度的可持續(xù)性模型初探[J].水文,2010,(1):35-38.(ZOU Jin.Discussion on the Sustainable Model for Long-term Optimal Operation of Reservoirs[J].Journal of China Hydrology,2010,(01):35-38.(in Chinese))
[4] 宋萌勃,岳延兵,陳吉琴.水庫調(diào)度與管理[M].鄭州:黃河水利出版社,2013.(SONG Meng-bo,YUE Yan-bing,CHEN Ji-qin.Reservoir Operation and Management[M].Zhenzhou:The Yellow River Water Conservancy Press,2013.(in Chinese))
[5] 何俊仕,林洪孝.水資源規(guī)劃及利用[M].北京:中國水利水電出版社,2006.(HE Jun-shi,LIN Hong-xiao.Water Resources Planning and Utilization[M].Beijing:China Water Power Press,2006.(in Chinese))
[HJ1.85mm][6] 王萬良,周慕遜,管秋,等.基于遺傳算法的小水電站優(yōu)化調(diào)度方法的研究與實(shí)踐[J].水力發(fā)電學(xué)報(bào),2005,24(3):6-11.(WANG Wan-liang,ZHOU Mu-xun,GUAN Qiu,et al.Research and Practice of Optimum Operation Method Based on Genetic Algorithm for Small Hydropower Stations[J].Journal of Hydroelectric Engineering,2005,24(3):6-11.(in Chinese))[JP]
[7] 徐剛,馬光文.基于蟻群算法的梯級水電站群優(yōu)化調(diào)度[J].水力發(fā)電學(xué)報(bào),2005,24(5):7-10.(XU Gang,MA Guang-wen.Optimal Operation of Cascade Hydropower Stations Based on Ant Colony Algorithm[J].Journal of Hydroelectric Engineering,2005,24(5):7-10.(in Chinese))
[8] 李崇浩,紀(jì)昌明,李文武.改進(jìn)微粒群算法及其在水庫優(yōu)化調(diào)度中的應(yīng)用[J].中國農(nóng)村水利水電,2006,(2):54-57.(LI Chong-hao,JI Chang-ming,LI Wen-wu.Modified Particle Swarm Algorithm and Its Application in Reservoir Operation Optimization[J].China Rural Water and Hydropower,2006,(2):54-57.(in Chinese))
[9] 梁旭,黃明.現(xiàn)代智能優(yōu)化混合算法及其應(yīng)用[M].北京:電子工業(yè)出版社.2011.(LIANG Xu,HUANG Ming.Modern Intelligent Optimization Algorithm and Application[M].Beijing:Publishing House of Electronics Industry,2011.(in Chinese))[ZK)]
[10] [ZK(#]Maurice Clerc.Particle Swarm Optimization [M].America:Wiley-ISTE,2006.
[11] N C Chauhan,M V Kartikeyan,A Mittal.A Modified Particle Swarm Optimizer and Its Application to the Design of Microwave Filters [J].Journal of Infrared,Millimeter and Terahertz Waves.2009,598-610.
[12] Eberhart R C,Shi Y.Particle Swarm Optimization:Developments,Applications and Resources[A].Proceedings of the IEEE Congress on Evolutionary Computation,Piscataway[C].NJ:IEEE Service Center,2001:81-86.
[13] 徐文星,耿志強(qiáng),朱群雄,等.基于SQP局部搜索的混沌粒子群優(yōu)化算法[J].控制與決策,2012,(4):557-561.(XU Wen-xing,GENG Zhi-qiang,ZHU Qun-xiong,et al.Chaos Particle Swarm Optimization Algorithm Integrated with Sequential Quadratic Programming Local Search [J].Control and Decision,2012,( 4):557-561.(in Chinese))
[14] 安曉會,高岳林.混合變異算子的自適應(yīng)粒子群優(yōu)化算法[J].計(jì)算機(jī)應(yīng)用,2008,(S1):28-30.(AN Xiao-hui,GAO Yue-lin.Adaptive Particle Swarm Algorithm with Hybrid Mutation Operator [J].Journal of Computer Applications,2008,(S1):28-30.(in Chinese))
[15] Maurice Clerc.The Swarm and the Queen:Towards a Deterministic and Adaptive Particle Swarm Optimization[C].Proceedings of International Conference on Evolutionary Computation,1999:51-57.
[16] 邱林,肖琳.改進(jìn)微粒群優(yōu)化算法在水庫防洪調(diào)度中的應(yīng)用[J].人民黃河,2007,29(10):18-21.(QIU Lin,XIAO Lin .Application of Improved Particle Swarm Ooptimization Algorithm in Reservoir Flood Control Operation[J].Yellow River,2007,29(10):18-21.(in Chinese))