基于強(qiáng)化學(xué)習(xí)協(xié)同進(jìn)化算法求解柔性作業(yè)車間節(jié)能調(diào)度問(wèn)題

打開文本圖片集

關(guān)鍵詞：柔性作業(yè)車間調(diào)度；Q-learning；改進(jìn)NSGA-II；多目標(biāo)優(yōu)化 中圖分類號(hào)：TH165 文獻(xiàn)標(biāo)志碼：A 文章編號(hào)：1001-3695（2025）07-016-2039-09 doi：10.19734/j. issn.1001-3695.2024.11.0479

Abstract：Fortheflexiblejob shopenergy eficientscheduling problem（EEFJSP），thispaperconstructedaFJSPmodel with theoptimizationobjectivesof minimizingthemaximumcompletiontimeand minimizingthetotal energyconsumption.Firstly， it proposedanadaptivealgorithmbasedonreinforcementlearningco-evolutionaryalgorithm（QNSGA-II）to characteriethe problemmodel．Secondly，it introduced theconceptsof state spaceandaction spaces，and designedareward-punishment functionbasedoheoverallaveragefinesandpopulationdiversitytoensuretheefectivenessofthealgoritinteiterative processInorder toimprovetheabilityof theglobal searchandlocalsearch，itproposedanimproved tabusearchalgorithmto updatethepopulationaftercrosoverandmutation.Inordertoimprovetheabilityof globalsearchandlocal search，itproposedanimprovedtaboosearchalgorithmtoupdate thepopulationaftercrosoverandmutation.Finall，itanalyzedtheeffectivenessoftheimproved tabusearch strategyandthe Q-learning parameteradaptationstrategy toverifythealgorithm’seffectiveness andsuperiority，anditcomparedtheproposedQNSGA-Iwithother multi-objectiveoptimizationalgorithmsto verify the superiority of the algorithms in solving the EEFJSP.

Key Words：flexible job shop scheduling；Q-learning；improved NSGA-I ；multi-objective optimization

0 引言

近幾十年來(lái)，隨著環(huán)境的不斷惡化，政府呼呼節(jié)能減排，綠色制造已成為各國(guó)增強(qiáng)制造業(yè)核心競(jìng)爭(zhēng)力的戰(zhàn)略重點(diǎn)。（剩余20115字）