資源描述:
《基于遺傳算法煉廠生產(chǎn)計(jì)劃優(yōu)化的研究》由會(huì)員上傳分享,免費(fèi)在線(xiàn)閱讀,更多相關(guān)內(nèi)容在學(xué)術(shù)論文-天天文庫(kù)。
1、摘要隨著石化企業(yè)管理水平的不斷提高,煉廠生產(chǎn)計(jì)劃優(yōu)化引起越來(lái)越多學(xué)者的重視,但目前大部分研究均集中于優(yōu)化模型的改進(jìn)和問(wèn)題規(guī)模的縮減,針對(duì)優(yōu)化算法的系統(tǒng)研究還不多見(jiàn)。因此本文在綜述煉廠生產(chǎn)計(jì)劃優(yōu)化的研究現(xiàn)狀后,對(duì)遺傳算法求解帶非線(xiàn)性和不確定性的優(yōu)化問(wèn)題進(jìn)行了嘗試,并在此基礎(chǔ)上設(shè)計(jì)了集成遺傳算法優(yōu)化平臺(tái),實(shí)現(xiàn)工廠級(jí)實(shí)例的求解.全文共由六章組成:第l章闡述了煉廠生產(chǎn)計(jì)劃的工業(yè)背景和基本概念,以及數(shù)值優(yōu)化算法的研究和應(yīng)用現(xiàn)狀.接著綜述了遺傳算法在該領(lǐng)域的應(yīng)用,并介紹了基于仿真優(yōu)化架構(gòu).針對(duì)當(dāng)前研究的不足,提煉本文的創(chuàng)新點(diǎn).第2章詳細(xì)介紹基本遺傳算法的思想、原理和步驟,并重點(diǎn)歸納典型的約束處理策略和
2、加速求解策略.進(jìn)一步,總結(jié)實(shí)驗(yàn)驗(yàn)證的遺傳算法參數(shù)設(shè)置經(jīng)驗(yàn),為后文的求解應(yīng)用提供了借鑒.第3章援引相關(guān)文獻(xiàn),建立石化生產(chǎn)中重點(diǎn)裝置的非線(xiàn)性經(jīng)驗(yàn)?zāi)P?,并?xì)化儲(chǔ)罐模型的描述,增加對(duì)應(yīng)的大量不等式約束.改進(jìn)優(yōu)化算法,構(gòu)建一種兩階段的約束處理策略,同時(shí)混合直接搜索方法提升了局部收斂速度。第4章對(duì)煉廠需求和產(chǎn)率的雙重不確定性條件進(jìn)行建模,采用基于仿真優(yōu)化架構(gòu).深入研究MonteCarlo仿真的合理采樣次數(shù),并引入馬爾可夫鏈的產(chǎn)率波動(dòng)模型.提出多種群并行求解的加速策略,優(yōu)化了算法的求解性能.第5章針對(duì)遺傳算法尚未有大規(guī)模工程應(yīng)用的現(xiàn)狀,進(jìn)行集成優(yōu)化平臺(tái)的開(kāi)發(fā).建立以現(xiàn)場(chǎng)數(shù)據(jù)為基礎(chǔ)的工廠模型,針對(duì)模型特點(diǎn)
3、設(shè)計(jì)以遺傳算法為核心的優(yōu)化子系統(tǒng).最后根據(jù)軟件質(zhì)量評(píng)價(jià)的關(guān)鍵指標(biāo)橫向?qū)Ρ绕渌麅?yōu)化軟件,說(shuō)明其應(yīng)用價(jià)值。第6章最后在總結(jié)全文的基礎(chǔ)上,展望了遺傳算法在煉廠生產(chǎn)計(jì)劃優(yōu)化應(yīng)用領(lǐng)域未來(lái)的發(fā)展方向.關(guān)鍵詞:石化企業(yè)生產(chǎn)計(jì)劃非線(xiàn)性不確定性遺傳算法約束處理加速求解集成優(yōu)化平臺(tái)111浙江大學(xué)碩士學(xué)位論文(此頁(yè)留空)ABSTRACTRefineryplanningoptimizationhasattractedmoreandmoreresearchers,withimprovementofpetrochemicalindustrymanagement.However,mostresearcherspaida
4、ttentioninmodeladjustmentanddimensionalityreductionwhileworkshasnotbeenreportedonsystematicresearchofoptimizationalgorithm.Therefore,somerelatedworkabouttheoptimizationofnonlinearanduncertainproblemsiscardedoutafterthemajorissuesonrefineryplanningoptimizationaresummarized.Meanwhile,anintegratedopt
5、imizationplatformisdeveloped.Themaincontentsofthisdissertationareasfollows:InChapter1,theoverviewofpetrochemicalindustryandthebasicconceptsofrefineryplanningareintroduced,aswellasthelatestresearchandapplicationofoptimizationalgorithminpetrochemicalindustry,especiallyapplicationof·geneticalgorithm.
6、What’Smore,asimulation-basedoptimizationframeworkisproposed.Later,thisthesis’SresearchobjectivesandinnovativepointsareInChapter2,thegeneticalgorithm’Stheoryissystematicallyintroduced,focusingonconstraintshandlingandacceleratingtechniques.Experiencesofparameterssettingarelistedinordertohelpthelater
7、solvingapplication.InChapter3,nonlinearmodelsofseveralimportantunitsarebuiltandtankmodelsaredescribedindetailSOthatlotsofinequalityconstraintsarebroughtin.Atwo-scaleconstraintshandlingstrategyisputforward勰wellash