Abstract: When using the SDO，there are always some problems，efficiency，consistency and so on.In order to solve the two problems by SDO while connecting the isomerous data source，an organon is raised which is grammar-oriented and automata theo-ry-oriented.The ontology is imported to describe the data exchange by semantic，and to enhance the SDO semantic analysis，and also a graphic structure is built up to access the isomerous data source by DMS.Because of the frequent alteration of the data graphic demands and the DAS，it’s required to cut out the data object to be the optimize trees，and then build up the mathematic matrix which would be optimized by the ant colony algorithm.At last，the policy mentioned above is used in the open source SDO API.It’s proved that the accessing to the data source has been improved a lot，and also the coherence and the semantics in data source are greatly enhanced too.

Key words: ontology, service data objects, data exchange, Service-Oriented Architecture（SOA）, automata, ant colony algorithm

摘要： 在使用服务数据对象SDO（Service Data Objects）过程中，针对SDO连接异构数据源存在的效率和一致性问题，提出一种面向文法和自动机的推理方法，同时，引入本体进行语义描述数据交换，增强SDO语义识别，并通过数据中介服务（DMS）建立一种异构数据源访问数据图的图结构，根据数据图的需求变更以及数据访问服务（DAS）实现行图裁剪成一棵棵数据对象最优树，建立相应的优化数学模型，并通过蚁群算法优化寻优该模型。最后通过在开源件SDO API中引入策略应用表示：提高了不同的数据源访问效率，增强了数据交换的一致性和语义性。

关键词: 本体, 服务数据对象, 数据交换, 面向服务的体系结构（SOA）, 自动机, 蚁群算法