摘要
Internet 环境下的知识表示与自动推理是广大人工智能及相关领域研究者面临的新课题。为了给自动推理领域提供一个统一的问题(知识)表示格式及为 Web 用户提供自动推理 Web 服务,提出了一个基于多种类逻辑的知识标记语言 MSKML(Many Sorted Knowledge Markup Language 的简写),实现了使用 RLD(Rightmost Linear Deduc-tion)演绎策略的多种类谓词演算的 COM 对象 Prover 来提供自动推理 Web 服务 ProverService。ProverService 采用简单对象访问协议 SOAP(Simple Object Access Protocol)作为服务器与 Web 客户间的网络通信协议,用 Web 服务描述语言 WSDL(Web Service Description Language)对其提供的服务进行了描述。最后,运用 RLD 演绎成功求解了著名的人工智能问题—Steamroller 问题,并以它为例,描述了该问题及 RLD 反驳的 MSKML 表示以及自动推理 Web 服务的调用过程。
The knowledge representation and automated reasoning on the Web are an important issue for artificial intelligence community. In order to provide automated reasoning community with a uniform format of knowledge representation, a many-sorted knowledge markup language （abbr. MSKML） is proposed. Furthermore, in order to provide Web users with automated reasoning Web service, a prototype system of many-sorted predicate calculus with RLD （Rightmost Linear Deduction）, i. e. , the COM object Prover, is implemented. The Prover has a Web service ProverService that uses the SOAP （Simple Object Access Protocol） as the communication protocol between the Prover and Web users. The information of ProverService is described using WSDL （Web Service Description Language）. When a client on the Web need the automated reasoning services, it represents the problem as a XML （Extensible Markup language） file conforming to the MSKML DTD （Document Type Definition） and sends a SOAP request message including the XML file to the automated reasoning server, and the automated reasoning server represents the solution as a XML file conforming to the MSKML DTD and send it back to the client as the SOAP response message. Finally, the wellknown artificial intelligence problem--the Steamroller problem--is successfully solved by Prover and the MSKML representations of the problem and RLD refutation are given to illustrate our work.
出处
《计算机科学》
CSCD
北大核心
2006年第5期205-210,共6页
Computer Science
基金
本工作得到广东省自然科学基金(编号020199)
华南理工大学自然科学青年基金的资助
