中文摘要：

针对代理发现中用户对代理的性能、成本和安全性等方面的需求,以及需求匹配过程中的隐私保护问题,基于Paillier同态加密算法,提出一种新的综合考虑代理和用户属性及其偏好的私有数据信息匹配算法,包括建立基于欧氏距离的相似度函数、利用加密算法进行匹配、计算相似度和确定匹配的代理链4个步骤。该算法引入半可信主代理从全局层面管理所有子代理的业务类型和连接状况,并承担主要的计算开销,同时将欧氏距离与Paillier同态加密算法有机结合,支持具有偏好信息的多元属性数据匹配,能够有效保障用户和子代理的安全性。最终,通过安全性分析与性能仿真,证明所提出方案的安全性和有效性。

英文摘要：

According to user＇s requirements on the proxy performance, cost and safety in proxy discovery, and the privacy-preserving issue during the process of the demand private matching, a new private matching algorithm was presented based on Paillier homomorphic encryption, comprehensively considered the attributes of user and agents and their priority. It included four steps： building the similarity function based on Euclidean distance, carrying on the private matching by encryption algorithm, calculating the similarity and screening the proxy chain. Proposed scheme introduced semi trusted primary proxy from the global level, which is to manage all the sub proxy＇s business type and the connection status, and do the main computational overhead.At the same time, Euclidean distance and Paillier homomorphic encryption algorithm were combined to support the multivariate attributes with priority to match, which can effectively protect the privacy of user and sub proxy. Finally, security analysis and evaluation results show the effectiveness and safety of proposed scheme.