云制造环境下基于本体和模糊QoS的供应商匹配方法  

Supplier Matching Method Based on Ontology and Fuzzy QoS in Cloud Manufacturing Environments

作　　者：孙晓琳 金淳 马琳 王文波[2] 

机构地区：[1]大连理工大学管理与经济学部,辽宁大连116024 [2]大连大学经济管理学院,辽宁大连116633

出　　处：《中国管理科学》2018年第1期128-138,共11页Chinese Journal of Management Science

基　　金：国家自然科学基金资助项目(71271041);辽宁省自然科学基金资助项目(2013020006)

摘　　要：云制造环境下的服务匹配具有资源数量大、语义信息不对称、QoS多样化和模糊化的特点,同时企业有自主选择匹配结果的需求。为此,提出基于本体和模糊QoS聚类的三阶段供应商匹配模型。首先构建本体模型和供应商服务描述模型,运用语义本体既消除了信息的不对称性,又增加了语义信息的完整性。此外,对QoS的多属性信息进行三角模糊化处理,结合模糊偏好和优化的模糊C均值聚类（FCM）算法按需聚类,提高了收敛速度和精度,得出基于匹配度排序的结果集合。实例验证结果表明：本文匹配方法较传统方法有更高的适应性和查准率。For the supplier selection of manufacturing enterprise in cloud manufacturing environment, a larger range of choices and the wide distribution of manufacturing resources are highly shared compared with the traditional manufacturing environment. Moreover, the fuzzy features of QoS bring new challenges for the supplier selection in cloud manufacturing environment. Therefore, large quantity of resource, semantic information asymmetry, and fuzzy of QoS become the key problems in supplier service matching. On a cloud manufacturing platform, the suppliers as services can be described by functional informa- tion and QoS information. Functional information is composed of concepts, numerical and interval. QoS information is represented by fuzzy language. Because of the large number of suppliers, functional information of supplier sl and supplier s2 are probably the same, but they almost have different QoS information. Accurate matching results can be obtained by matching the two kinds of information in two services. In this paper, a three-phase service matching model is proposed based on ontology and fuzzy QoS clus- tering. Firstly, a description models of service description and ontology is established with semantic ontology in order to eliminate the asymmetry of information and increase the integrity of the semantic information. Secondly, the multiple attributes of QoS based on the triangular fuzzy number are established by combine with fuzzy preference and optimize fuzzy c-means clustering algorithm （FCM）, greatly improve the speed and efficiency of convergence. Finally, the experiment is conducted according to real automobile supplier data and expert opinions, and the results from the actual experiment have shown that this method can achieve higher precision and adaptability compared with the traditional methods. In this study, new idea,whinch is about how to solve the problem of service matching in cloud manufacturing environment is put forward.

关 键 词：服务匹配 供应商服务本体 模糊QoS 模糊C均值聚类算法 聚类中心 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]