Vol.27 No.6 鄢旭等：基于超立方体与信息熵的离散化方法 ·763· 因此，计算属性的复杂度由原来的O(4)降低 北京科技大学学报，2005,273：364 为现在的O(4Y++YP):同时对断点重要性 [3)王国胤.Rough集理论与知识获取.西安：西安交通大学出 版社，2003 进行了量化，因此对断点的选取更为合理，能够 [4)苗夺谦.Rough Set理论中连续属性的离散化方法.自动化 产生较小数目的断点集；引入分类信息熵作为约 学报，200L,27(3:296 束条件，保证了断点选取得方向性和可靠性，同 [5]Nguyen H S,Skowron A.Boolean reasoning for feature extrac- 时进一步起到了属性约简、属性值约简的作用， tion problems.In:10th International Symposium on Foundations 降低了离散算法的时间复杂度和空间复杂度， of Intelligent Systems.New York:Springer-Verlag,1997.116 [6]Nguyen S H,Skowron A.Quantization of real value attributes- rough set and boolean reasoning approach.Bull Int Rough Set 参考文献 Soc,1996(1):347 [1]Pawlak Z.Rough Set.Int J Comput Inf Sci,1982 (1):341 [7)孟庆生.信息论.西安：西安交通大学出版社，1986 [2]鄂旭，高学东，武森，等.信息表中不完备数据的填补方法 Discretization algorithm based on super-cube and information entropy EXU,GAO Xuedong",TAN Wendong",WANG Ying" 1)Department of Computer Science,Liaoning Institute of Technology,Jinzhou 121001,China 2)Management School,University of Science and Technology Beijing,Beijing 100083,China ABSTRACT Discretizing continuous attributes in a rough set were researched.Based on the concept of super-cube, all attributes of the information table in data space were globally discretized.By the consistent correlation of condi- tion attributes and decision attributes,important condition attributes were selected depending on their classifying ability in the rough set boundary zone,and furthermore,important breaking points were selected to discretize the in- formation table on a single attribute locally with the iterative constraints of information entropy.Illustration and ex- perimental results indicate that the algorithm combining the global and local discretization is effective and efficient. KEY WORDS rough set;discretization;super-cube;information entropy 表本冷茶冷李素条率者米条米*杂孝卷幸冷本为*米条年冬海奇为女★*染为*米激余孝章装条次餐女米中冷米条弹京备整个女*资*★路今率京食薄海冷★*米**米冷★率郑素余冷*冷★冷数章女 (上接第739页) Production mode optimization of a steelmaking workshop system LIU Oing",HUANG Xingwu,FU Pingyuan 1)Metallurgical and Ecological Engineering School,University of Science and Technology Beijing,Beijing 100083,China 2)Xinjiang Bayi Iron Steel Co Ltd.Urumchi 830022,China 3)Linfen Iron and Steel Co.Ltd.,TISCO Group,Lifen 041000,China ABSTRACT The problem about running mode optimization of a steelmaking workshop system,which is the most complicated subsystem in a steel manufacturing process system was discussed by applying the theory of metallurg- ical engineering.The level and logical relation of the running mode of a steelmaking workshop in the productive process were described and studied in combination with its applications in Chinese four BOF workshops with the technology of high efficiency continuous casting.Three effects of production mode optimization were summarized and illustrated in detail:(1)the layout of a workshop tending to rationality and briefness,(2)the buffer of working procedures tending to rationality and efficiency,and(3)the operation of working procedures tending to high effi- ciency and optimization.These conclusions are applied to the research of optimization and control in special steel- making process operation,and valuable results are acquired. KEY WORDS steelmaking workshop;ferrous metallurgy process engineering;production mode;process optimi- zation:mass flow、勺 l 一 7 N o 2 一 鄂旭 等 基于 超立方 体 与信 息嫡 的离 散 化方 法 : 6 . 7 6 3 - 因 此 , 计算 属 性 的复 杂度 由原 来 的 O (l川}酬 2 ) 降低 为现 在 的口(团( 1艺1 2 +. 二 + {玖 } , ) ; 同时对 断 点重 要 性 进行 了量化 , 因 此对 断 点 的选取 更 为 合理 , 能够 产生 较小数 目的 断点集 ; 引入 分类 信息 嫡 作为 约 束条 件 , 保 证 了断点 选取 得 方 向性 和 可靠 性 , 同 时进 一 步起 到 了属 性 约简 、 属 性值 约 简 的作 用 , 降低 了离散算 法 的时 间复 杂度 和 空 间 复杂 度 . 参 考 文 献 f l ] P a w lak 2 . OR u g h S e t . I n t J C o 二 p u t I n f s e i , 19 8 2 ( l ) : 34 1 2[ ] 鄂旭 , 高学 东 , 武森 , 等 . 信 息表 中不完 备数据 的填 补方 法 北京 科技 大学 学报 , 2 0 0 5 , 2 7( 3 ) : 36 4 3[] 王 国溉 . oR u gh 集 理论 与知识 获取 . 西 安 : 西安 交通 大学 出 版 社 , 2 0 0 3 4[] 苗 夺谦 . oR gu h se t 理 论 中连 续 属性 的离散 化方 法 . 自动化 学报 , 2 00 1 , 2 7 ( 3 ) : 29 6 [5 1 N g u界n H S , S k o w or n A . B o o l e an 祀 a s o n in g of r fe a t u r e e xtr a e - t i o n P or b l e m s . I n : 10 ht I n t e l刀 a t i o n a l S y m P o s i um o n F o un dat i on s o f l n t e llig e n t s y s t e m s . N e w 丫b r ` : S P ir n g e -r Ve r l a g , 1 99 7 . 116 [6 ] N g ly e n S H , S k o w or n A . Q aun t i z at ion o f re a l v a lue a itr b u t e s - or u g h s e t an d b o o l e a n fe a s o n i n g ap pm ac h . a u u I n t OR u gh S e t S o c , 19 9 6 ( 1 ) : 34 7 7[ 」 孟庆生 . 信 息论 . 西 安 : 西安 交通 大学 出版社 , 19 86 D i s e re ti z at i o n a lg o r it h m b a s e d o n s uP e -r c u b e an d i n of rm a t i o n e n otr P y E X zj l只 GA 口 xu e do 心气TA N 确 n do gn ,气恻刃 G 五n g Z ) 1) D e P a rt m e n t o f C o mP ut e r s e i e n e e , L iao n i n g l n s t itU t e o f eT e hn o l o g又 Jinz ho u 12 1 00 1 , C h i n a 2 ) M a n a g e m e nt s e h o o l , nU iv e r s ity 0 1 ’ s e i e n e e an d eT e hn o l o g y B e ij i n g , B e ij i雌 10 0 0 8 3 , C h i n a A B S T R A C T D i s e re t i z i n g c o n t i n u o u s a t r i b ut e s i n a or u g h s e t w e re re s e 毗he d B a s e d o n ht e e o cn e Pt o f s uP e -r c ub e , a ll a t r lb ut e s o f ht e in fo mr at i o n abt l e i n d at s P a c e w e er g l o b a ll y d i s e er t i z e d . B y ht e e o n s i s te in e o 讹 l at i o n o f e o n di - t i o n a itr b ut e s an d d e e i s i o n 川itr b u t e s , im P o rt a l l t e o n d i ti o n a t r i b u t e s w e er s e l e e et d d e P e n d in g o n ht e i r e l a s s ify i n g ab i l i yt i n ht e or u g h s et b o u n d a yr z o n e , a n d fu rt h e mr o er , im P o rt a n t b er ak i n g P o int s w e代 s e l e e t e d ot d i s e er ti z e ht e i n - fo mr at ion t ab l e o n a s i n g l e at r i b ut e l o e a l ly w i ht ht e it e r at i v e e o n s tr a i n t s o f i n fo mr at i o n e n t or P y . Il l u s tr at i o n an d e x - P e ir m e n t a l er s u lt s i n d i c at e ht a t ht e a l g o ir t h m e o m b i n i n g ht e g l o b a l an d l o c a l d i s e er t i z at i o n 1 5 e fe e t i v e an d e if e i e n t . K E Y WO R D S r o u g h s e t : d i s; e er t i z a t i o n : s uP e -r e u b e : i n fo mr at i o n e nt or P y (上 接第 7 39 页 ) P or d u c it o n m o de o P t ixn i z at i o n o f a s et e lm a k i n g w o kr s h o P s y s t e m “ “ Qi n g , ), H。月刃 G ix 刀 g w: 。, ), F U p i刀 g夕u a n , , l ) M e at ll u 电 i e a l a dn E e o l o g i e a 1 E n g i r Le e ir gn S e h o o l , U n i v e r s ity o f s e i e n c e an d eT e h n o l o g y B e ij i n g , B e ij i n g l 0 0 0 83 , C h i n a 2) X inj i a n g B ay i rID n & S t e e l C o L t d . U rum e h i 8 3 00 2 2 , C h i n a 3) L i n fe n Ior n a n d S t e e l C o . L td , T I S C O G or uP , L ife n 04 1 00 0 , C h i n a A B S T R A C T T h e Por b l e m al ) o u t run n i n g m o d e o P t im i z at i o n o f a s t e e lm ak i n g w o rk s h o P s y s te m , w h i e h 1 5 ht e m o s t e o m P l i e at e d s u b s y s t e m i n a s t e e l m an u af e tUr i n g P or e e s s s y s t e m w a s d i s e u s s e d b y ap P ly in g ht e ht e o yr o f m e t a lin gr - i e a l e n g i n e e ir n g . T h e l e v e l ar 一d l o g i e a l er lat i o n o f ht e unr n i n g m o d e o f a s t e e lm a ik n g w o kr s h o P i n ht e Por du e t i v e P or e e s s w er d e s e r ib e d an d s l u d i e d i n c o m bi n at i o n w ith it s ap P li e at i o n s i n C h i n e s e fo ur B O F w o kr s ho P s w iht ht e te e hn o l o gy o f h i g h e if e i e n c y e o n t i n u o u s e a s t i n g . T h er e e fe c t s o f P or du e t i o n m o d e o P t im i z at i o n w e er s u n l n l ar i z e d an d i l l u s tr at e d i n d e t a i l : ( l ) ht e l盯o u t o f a w o kr s h o P t e n d i n g t o r a t i o n a liyt an d b ir e hf e s s , ( 2 ) ht e b u fe r o f w o kr i n g P or e e d u er s t e n d i n g t o art i o n a j i yt an d e if e i e n e y, an d ( 3 ) ht e o P e art i o n o f w o kr i n g orP e e d u er s t e n d i n g t o h ihg e if - e i e n e y an d o Pt im i z at i o n . hT e s; e e o n e fu s i o n s aer a P P li e d t o ht e er s e娥h o f o P tim i z at i o n an d e o n tDI I i n s P e e i a l s t e e l - m ak i n g Por e e s s o P e r at i o n , an (1 v a l u ab l e er s u l t s aer a e q u i er d . K E Y W O R D S st e e lm iak n g w o r k hs op : fe our s m e at ll u r g y P r o e e s s e n g i n e e r i n g : P or du e it o n m o d e : Por e e s s o P t im i - z at i o n ; m a s s fl o w