北京科技大学学报
北京科技大學學報
북경과기대학학보
JOURNAL OF UNIVERSITY OF SCIENCE AND TECHNOLOGY BEIJING
2014年
1期
123-130
,共8页
董广静%施灿涛%李铁克%王柏琳
董廣靜%施燦濤%李鐵剋%王柏琳
동엄정%시찬도%리철극%왕백림
钢管%堆垛%聚类算法%约束满足问题%决策
鋼管%堆垛%聚類算法%約束滿足問題%決策
강관%퇴타%취류산법%약속만족문제%결책
steel tubes%stacking%clustering algorithms%constraint satisfaction problems%decision making
针对钢管入库优化决策问题,建立了问题的约束满足优化模型,并通过对垛高和钢管堆放规则的分析,提出了基于聚类和约束满足技术的两阶段求解算法。算法在第一阶段采用聚类的方式对待入库的钢管按照多重属性进行分组;在第二阶段利用约束满足技术对于每组钢管分别指派垛位及其在垛位上的具体位置,并通过约束传播动态缩减问题的搜索空间。最后将算法与经典的BFD ( best fit deceasing)算法进行实验结果对比。实验结果表明,算法能够在保证倒垛次数最小的前提下,有效减少垛位数并具有良好的垛位利用率,模型及算法可行、有效。
針對鋼管入庫優化決策問題,建立瞭問題的約束滿足優化模型,併通過對垛高和鋼管堆放規則的分析,提齣瞭基于聚類和約束滿足技術的兩階段求解算法。算法在第一階段採用聚類的方式對待入庫的鋼管按照多重屬性進行分組;在第二階段利用約束滿足技術對于每組鋼管分彆指派垛位及其在垛位上的具體位置,併通過約束傳播動態縮減問題的搜索空間。最後將算法與經典的BFD ( best fit deceasing)算法進行實驗結果對比。實驗結果錶明,算法能夠在保證倒垛次數最小的前提下,有效減少垛位數併具有良好的垛位利用率,模型及算法可行、有效。
침대강관입고우화결책문제,건립료문제적약속만족우화모형,병통과대타고화강관퇴방규칙적분석,제출료기우취류화약속만족기술적량계단구해산법。산법재제일계단채용취류적방식대대입고적강관안조다중속성진행분조;재제이계단이용약속만족기술대우매조강관분별지파타위급기재타위상적구체위치,병통과약속전파동태축감문제적수색공간。최후장산법여경전적BFD ( best fit deceasing)산법진행실험결과대비。실험결과표명,산법능구재보증도타차수최소적전제하,유효감소타위수병구유량호적타위이용솔,모형급산법가행、유효。
A constraint satisfaction optimization model was presented to deal with the optimization decision problem about the steel tube location. Through the analysis of stack height and the piling rules of steel tubes, a two-stage algorithm was given based on cluste-ring and constraint satisfaction technology. In the first stage, steel tubes to be put into storage are grouped by clustering-based approach according to their multiple attributes. In the second stage, by using constraint satisfaction technology, the specific location of steel tubes in each group is assigned, and the search space of the problem is dynamically shrunk through constraint propagation. Finally, this algorithm was compared with the classical BFD ( best fit deceasing) algorithm through experiments. Experimental results demon-strate that, in the premise of minimizing stacking operations, the algorithm can effectively reduce the quantity of stacks and achieve a well-performed utilization rate of stacks. And thus the results verify the feasibility and effectiveness of the model and algorithm.