(e) Die type   (f) Fixation  mode (g) Size tolerance (mm) (h) Ⅰ  0 0 1 1 0 0 0 1 Ⅱ  0 1 0 0 2 0 0 0 Ⅲ  1 2 1 1 1 1 1 1 Ⅳ  0 1 1 0 2 2 1 2 Ⅴ  1 1 1 0 2 2 0 2  U/Ind（D）={{Ⅰ},{Ⅱ},{Ⅲ},{Ⅳ},{Ⅴ}} U/Ind（abce）={{Ⅰ},{Ⅱ},{Ⅲ},{Ⅳ},{Ⅴ}}   U/Ind（abcd）={{Ⅰ},{Ⅱ},{Ⅲ},{Ⅳ},{Ⅴ}} U/Ind（aced）={{Ⅰ},{Ⅱ},{Ⅲ},{Ⅳ},{Ⅴ}}  U/Ind（bcde）={{Ⅰ},{Ⅱ},{Ⅲ},{Ⅳ,Ⅴ}}  U/Ind（abde）={{Ⅰ},{Ⅱ,Ⅳ},{Ⅲ}{Ⅴ}}  pos C ( D)={ Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ} r C ( D) = card ( pos C ( D) ) / card (U)=5/5=1 pos e C− ( D) ={Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ}  r e C− ( D) = card ( pos e C− ( D) ) / card (U) =5/5=1 r C ( D)－r e C− ( D)=1－1=0  pos d C− ( D) ={Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ}  r d C− ( D) = card ( pos d C− ( D) ) / card (U) =5/5=1 r C ( D)－r d C− ( D)=1－1=0 pos c C− ( D) ={Ⅰ,Ⅲ,Ⅴ} r c C− ( D) = card (pos c C− ( D) ) / card (U) =6/5=0.6 r C ( D)－r c C− ( D)=1－0.6=0.4 pos b C− ( D) ={Ⅰ,Ⅱ,Ⅲ,Ⅳ,Ⅴ} r b C− ( D) = card (pos b C− ( D) ) / card (U) =5/5=1 r C ( D)－r b C−  ( D)=1－1=0 pos a C− ( D) ={Ⅰ,Ⅱ,Ⅲ} r a C− ( D) = card (pos a C− ( D) ) / card (U) =3/5=0.6  r C ( D)－r a C− ( D)=1－0.6=0.4 According to the characteristic attribute calculation of die case, we obtain the calculation result. The result is analyzed, the attributes of a and c are important to select die case. And we can improve the set of {a,c} which is the least reduction of the attribute table to establish the index. When we calculate the similarity degree, we only consider case characteristic attribute  of {a,c}. According to the calculation formula: n = k = l = 5,  1 β = 2 β = 0.4.Surpose the new design problem of die case to be described as: (Punch, 6.2, 380, 0.44t, Low carbon steel). According to the index, we select case Ⅰand case Ⅱwhich are regarded as reference case. Remove the attribute of {b, d, e} and according to the formula (3) and (4), we obtain: Case Ⅰ = 0.4 (1 + 380/ 410) = 0.7707      Case Ⅱ = 0. 4 (1 + 340/380) = 0.7579 It is obvious that case Ⅰ is close to the design object die. So we use the interrelated information of case Ⅰ which is regarded as the reference of design die.  6. Conclusions  The paper analyzes case representation method and retrieval strategy using RST sand CBR. It puts forward a method using grade classification and decision attributes support degree to deal with the quantitative characteristics. And it confirms the important degree of all types of characteristic attributes. It aims to build up a retrieval based on case's key attributes. Then it makes use of the nearest neighbor strategy to implement the similar matching between the target case and source case. The technology guarantees the validity of case retrieval reduces system dependence and improves efficiency of case retrieval.  References  [1]  Pawlak Z (1982). “Rough sets”.  International Journal of Computer and Information Sciences.Vol.11,No.5,pp 341-356 [2] 
Zhang Wenxiu, Qiu Guofang (2006). “Uncertain Decision Making Based on Rough Sets”.  Beijing: Tsinghua University Press. [3]  Masahiro Inuiguchi(2006). “Attribute Reduction In Variable Precision Rough Set Model”.International Journal of Uncertainty, Fuzziness and Knowledge-Based Systems, Vol.14,No.4,pp461-479 [4]  YANG Zhi, HU Jinzhu,HU Longjiang(2005). “Component Retrieval Based on Knowledge Base and Case Reasoning”. Computer Engineering, Vol.21,No.11,pp159-162 [5]  Albert Fornells, Elisabet Golobardes, Josep Maria Martorell, etc(2007). “A Methodology for Analyzing Case Retrieval from a Clustered Case Memory”.  Case-Based Reasoning Research and Development. No.8,pp122-136 [6]  D. Alisantoso, L.P. Khoo, B.H. Ivan Lee and S.C. Fok(2004). “A rough set approach to design concept analysis in a design chain”.The International Journal of Advanced Manufacturing Technology. No.12,pp427-435 [7]  Chuang-Cheng Chiu and Chieh-Yuan Tsai(2007). “A Weighted Feature C-Means Clustering Algorithm for Case Indexing and Retrieval in Cased-Based Reasoning”.  New Trends in Applied Artificial Intelligence. No.7,pp 541-551 [8]  SUN Yan-sheng; YUAN Fu-yu; YU Zhuo-er,etc(2007). “Extraction of Decision Rules Based on Rough Set and Evidence Theory”.  Journal of Jilin University, Vol.45,No.4,pp577-581 [9]  JIN Wei; CHEN Huiping(2007). “A Hybrid Hierarchical k-means Clustering Algorithm”. Journal of Hohai University Changzhou, Vol.21,No.3,pp7-10 [10] Wang S K M,Ziarko W(1985). “On option decision rules in decision tables”.Bulletin of Polish Academy of Science, No.33,pp693-696. [11] WANG Guo-yin(2003). “Calculation Methods for Core Attributes of Decision Table”.  Chinese Journal of Computers,Vol.26,No.5,pp611-615 [12] Tony Mileman, Brian Knight, Miltos Petridis , Don Cowell and J. Ewer (2006). “Case-based retrieval of 3-dimensional shapes for the design of metal castings”.Journal of Intelligent Manufacturing.No.11, pp39-45 [13] Wang Ke, Liao Wenhe, Guo Yu, etc (2007). “Research on Multi-layer Case Retrieval Model Based on Bitwise Indexing”.China Mechanical Engineering. Vol.18, No. 16,pp1953-1956 [14] Gao Changqing, Huang Kezheng, Zhao Fang, etc (2007). “Research on Fast Innovation Design Method of Structure Based on CBR”.China Mechanical Engineering. Vol.18, No.24,pp2907-2913 案例检索算法冲压模具英文文献和中文翻译(4):http://www.youerw.com/fanyi/lunwen_33959.html