中国机械工程
中國機械工程
중국궤계공정
China Mechanical Engineering
2015年
21期
2873-2879,2884
,共8页
邹攀%李蓓智%杨建国%施烁%梁越昇
鄒攀%李蓓智%楊建國%施爍%樑越昇
추반%리배지%양건국%시삭%량월승
柔性作业车间%智能调度%多目标%调度资源信息
柔性作業車間%智能調度%多目標%調度資源信息
유성작업차간%지능조도%다목표%조도자원신식
flexible job shop%intelligent scheduling%multi??objective%real??time resource informaG tion
针对离散制造柔性作业车间实际工况,提出了一种基于分层蚁群遗传算法的柔性作业车间资源驱动的多目标调度方法,其基本特征是:基于连续生产中不同调度周期剩余或空闲资源等调度相关实时信息;基于完工时间和机床负荷等多目标;采用分层蚁群遗传混合算法进行决策,通过逐步筛选,获得优化解.该方法特别适用于车间资源变化、任务执行情况变化、急件任务必须插入等情况下的动态调度.应用标准案例并设计相关组合案例进行了测试,与 MOGV 混合算法相比,25%的案例计算结果优于 MOGV 算法,最大完工时间减少5%~7%,62.5%的案例计算结果等同 MOGV 算法.因此,该智能调度方法不仅可以有效地取得对指定优先目标的最佳优化效果,且可自动获得多目标综合的最优解,智能调度效果显著.
針對離散製造柔性作業車間實際工況,提齣瞭一種基于分層蟻群遺傳算法的柔性作業車間資源驅動的多目標調度方法,其基本特徵是:基于連續生產中不同調度週期剩餘或空閒資源等調度相關實時信息;基于完工時間和機床負荷等多目標;採用分層蟻群遺傳混閤算法進行決策,通過逐步篩選,穫得優化解.該方法特彆適用于車間資源變化、任務執行情況變化、急件任務必鬚插入等情況下的動態調度.應用標準案例併設計相關組閤案例進行瞭測試,與 MOGV 混閤算法相比,25%的案例計算結果優于 MOGV 算法,最大完工時間減少5%~7%,62.5%的案例計算結果等同 MOGV 算法.因此,該智能調度方法不僅可以有效地取得對指定優先目標的最佳優化效果,且可自動穫得多目標綜閤的最優解,智能調度效果顯著.
침대리산제조유성작업차간실제공황,제출료일충기우분층의군유전산법적유성작업차간자원구동적다목표조도방법,기기본특정시:기우련속생산중불동조도주기잉여혹공한자원등조도상관실시신식;기우완공시간화궤상부하등다목표;채용분층의군유전혼합산법진행결책,통과축보사선,획득우화해.해방법특별괄용우차간자원변화、임무집행정황변화、급건임무필수삽입등정황하적동태조도.응용표준안례병설계상관조합안례진행료측시,여 MOGV 혼합산법상비,25%적안례계산결과우우 MOGV 산법,최대완공시간감소5%~7%,62.5%적안례계산결과등동 MOGV 산법.인차,해지능조도방법불부가이유효지취득대지정우선목표적최가우화효과,차가자동획득다목표종합적최우해,지능조도효과현저.
A hierarchical ant??genetic algorithm??based multi??objective intelligent scheduling algoG rithm was proposed for flexible job shop problem.Its basic features were:(1)the approach was based on the real??time resource information of different scheduling periods;(2)its targets were completion time and machine load etc.;(3)the multi??objective optimization strategy and method were used in an ant??genetic hybrid algorithm to obtain the optimal solution.This method could be used in the periodiG cal normal scheduling,the dynamic scheduling scenario and the situation of urgent jobs inserting. Some tests were done on the standard cases and a combined case.Compared to MOGV hybrid algoG rithm,the proposed approach outperformed in 25% of the test cases with a 5%~7% decrease in comG pletion time.As for rests 75% of test cases,the above two algorithms show the same results.ThereG fore,with the ability of optimizing results based on the priorities of objectives and the comprehensive performance of all objective automatically,the effectiveness of the method proposed in this paper was verified.