中文摘要：

针对如何高效地构建用于语义交互的分布式本体库展开了研究。首先，分析了基于本体的语义交互平台的通用体系结构，进而明确分布式本体库在平台中的作用，提出了一种分布式本体库存储模型。围绕该模型，详细讨论了网络本体库与本地本体库的映射方法，语义数据的编码方法，然后详细分析了该模型中指令分发和数据演化组件的工作机制，最后通过实验验证了：在集群存储模式下，对于大数据量查询，网络数据传输对查询时间开销的影响远远小于查询本身，从而为分布式本体库存储效率的优化奠定了基础，即在本体库中采用分布式集群存储可以达到优化检索效率的目的。该存储模型已经在STASIS平台中得到应用。

英文摘要：

Oriented to how to establish distributed ontology repositories efficiently to support semantic interaction among different users though the Internet, at the beginning, the paper reviews some semantic interoperation platforms based on ontology repositories to extract the common architecture of this kind of system, which helps to define the role of distributed ontology repositories in the whole system. A storage model for distributed ontology repositories is presented. According to the model, the method for making mapping between network ontology repositories and local repositories is discussed, and the encoding mechanism for semantic entities is explained. Then, the principles of the instruction dissemination controller and the data evolution controller are analyzed in detail. Finally, based on the experiment, it is validated that data transfer in network contributes much less to time cost than query itself in the cluster storage pattern, which presents the base for optimizing the retrieval efficiency in distributed ontologies repositories, namely, the optimization for the efficiency of data retrieval in ontologies repositories can be achieved by adopting the distributed cluster storage. The storage model has been applied to the project of the software for ambient semantic interoperable service （STASIS）.