Abstract: n the process of implementing Enterprise Service Architecture（ESA） software，using enterprise reference model effectively can accelerate enterprise modeling rate，improve software’s implementing efficiency，and reduce the implementation cost greatly.How to get the proper reference model form the reference model database is an important problem of reference model management.In view of the shortcomings of present popular reference model classification and retrieval algorithms，a reference model classification and retrieval algorithm based on rough set theory was put forward.The reference model’s characteristics were analyzed，and the characteristics were divided into two types：Fuzzy characteristics and specific characteristics.After the fuzzy characteristics being discreted，the reference model’s characteristics were treated using rough set theory，then the reference model which has maximum similarity with the newly building model was selected.At last，the correctness and effectiveness of this algorithm was attested through a case.

Key words: enterprise modeling, reference model classification and retrieval, rough set theory

摘要： 在实施企业服务框架软件的过程中，有效地利用企业参考模型可以加速建模进度，提高软件的实施效率，从而大大降低实施成本。如何从参考模型库中检索到合适的参考模型是参考模型库管理的一项重要内容。鉴于目前通用的分类检索算法存在的问题，提出了基于粗糙集的参考模型分类检索算法。分析了参考模型的特征，将参考模型的特征分为两类：模糊特征和确定特征。将模糊特征离散化后，采用粗糙集理论对参考模型的特征进行处理，检索到与在建模型相似度最大的参考模型。最后，通过应用实例验证了算法的正确性和有效性。

关键词: 企业建模, 参考模型分类检索, 粗糙集理论