求解水庫(kù)優(yōu)化調(diào)度問(wèn)題的動(dòng)態(tài)規(guī)劃_遺傳算法

《求解水庫(kù)優(yōu)化調(diào)度問(wèn)題的動(dòng)態(tài)規(guī)劃_遺傳算法》由會(huì)員上傳分享，免費(fèi)在線閱讀，更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。

1、第40卷第5期武漢大學(xué)學(xué)報(bào)(工學(xué)版)Vol.40No.52007年10月EngineeringJournalofWuhanUniversityOct.2007文章編號(hào):167128844(2007)0520001206求解水庫(kù)優(yōu)化調(diào)度問(wèn)題的動(dòng)態(tài)規(guī)劃2遺傳算法劉攀,郭生練,雒征,劉心愿(武漢大學(xué)水資源與水電工程科學(xué)國(guó)家重點(diǎn)實(shí)驗(yàn)室,湖北武漢430072)摘要:綜合動(dòng)態(tài)規(guī)劃和遺傳算法的優(yōu)點(diǎn),基于貝爾曼最優(yōu)化原理將水庫(kù)優(yōu)化調(diào)度按階段劃分為若干多目標(biāo)決策子問(wèn)題,各子問(wèn)題采用混合編碼的多目標(biāo)遺傳算法求解,從而提出了一種求解水庫(kù)優(yōu)化調(diào)度問(wèn)題的動(dòng)態(tài)規(guī)劃2遺傳算法.該算法不僅在時(shí)間上通過(guò)劃分階段降維,而且采

2、用遺傳算法克服離散狀態(tài)空間組合所產(chǎn)生的維數(shù)災(zāi)問(wèn)題.從理論上論證了動(dòng)態(tài)規(guī)劃2遺傳算法的全局收斂性,分析得出該算法的效率一般高于遺傳算法;并經(jīng)數(shù)值試驗(yàn)表明:在計(jì)算時(shí)段數(shù)較多時(shí),動(dòng)態(tài)規(guī)劃2遺傳算法顯著優(yōu)于遺傳算法.因此,提出的動(dòng)態(tài)規(guī)劃2遺傳算法為求解水庫(kù)優(yōu)化調(diào)度問(wèn)題提供了新的思路與途徑.關(guān)鍵詞:水庫(kù)調(diào)度;遺傳算法;最優(yōu)化原理;動(dòng)態(tài)規(guī)劃2遺傳算法中圖分類號(hào):TV697.1文獻(xiàn)標(biāo)志碼:AOptimizationofreservoiroperationbyusingdynamicprogramming2geneticalgorithmLIUPan,GUOShenglian,LUOZheng,LIUXi

3、nyuan(StateKeyLaboratoryofWaterResourcesandHydropowerEngineeringScience,WuhanUniversity,Wuhan430072,China)Abstract:BasedontheBellman’soptimalityprinciple,adynamicprogramming2geneticalgorithm(DP2GA)isproposedtooptimizethereservoiroperation.Thereservoiroptimaloperationispartitionedintomulti2stagepro

4、blemsbasedontheoptimalityprinciple,whichcanbeoptimizedbyusingmulti2objectiveGArespectively.Thedimensionsofbothtemporalandspatialcouldbereducedgreatly;andtheglobalconvergenceandefficiencyoftheDP2GAareanalyzedtheoretically.AMonteCarloexperimentisusedtosimulateandevaluatetheproposedalgorithmwiththeQi

5、ngjiangRivercascadehydropowerstationsascasestudy.ItisindicatedthattheDP2GAisvaluableforthereservoiroptimaloperation,especiallyfortheoptimizationofreservoiroperationwithmanyperiods.Keywords:reservoiroperation;geneticalgorithm;Bellman’soptimalityprinciple;dynamicprogram2ming2geneticalgorithm尋求水庫(kù)最優(yōu)調(diào)度

6、軌跡過(guò)程線,是水庫(kù)優(yōu)化調(diào)方法,在水庫(kù)(群)優(yōu)化調(diào)度領(lǐng)域中得到了較多的關(guān)度中的經(jīng)典、難點(diǎn)問(wèn)題.動(dòng)態(tài)規(guī)劃是水庫(kù)調(diào)度中應(yīng)注,可在一定程度上克服維數(shù)災(zāi),給現(xiàn)代水庫(kù)調(diào)度[3～7]用最廣泛的方法,它對(duì)優(yōu)化問(wèn)題從時(shí)間上降維,但技術(shù)注入了活力與生機(jī).遺傳算法雖能收斂于[1,2]仍避免不了狀態(tài)空間的維數(shù)災(zāi)問(wèn)題.遺傳算法全局最優(yōu)解,但需耗用相當(dāng)長(zhǎng)的計(jì)算時(shí)間作為代是一種基于模擬自然基因和自然選擇機(jī)制的尋優(yōu)價(jià),因此計(jì)算時(shí)間是遺傳算法應(yīng)用中的關(guān)鍵問(wèn)收稿日期:2006211229作者簡(jiǎn)介:劉攀(19782),男,湖南湘潭人,博士,講師,從事水文及水資源開(kāi)發(fā)利用研究.基金項(xiàng)目:國(guó)家自然科學(xué)基金(編號(hào):50609017)

7、;湖北省自然科學(xué)基金(編號(hào):2006ABA229).?1994-2010ChinaAcademicJournalElectronicPublishingHouse.Allrightsreserved.http://www.cnki.net2武漢大學(xué)學(xué)報(bào)(工學(xué)版)2007[3]k+1kk+1k題.如采用1年的資料進(jìn)行單一水庫(kù)優(yōu)化調(diào)度,水量平衡方程:QIi=Qoi+Qi;其中QIi,Qoi,Qi計(jì)算時(shí)段長(zhǎng)度為月(優(yōu)化變量個(gè)數(shù)