Object-Oriented Knowledge Base Management Systems

面向对象的知识库管理系统

• 批准号: 8810292
• 负责人: Phillip Sheu
• 金额: $ 1.36万
• 依托单位: Purdue Research Foundation
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 1988
• 资助国家: 美国
• 起止时间: 1988-06-15 至 1989-03-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8810292&HistoricalAwards=false
• 关键词: Object; Oriented; Knowledge; Base; Management

In the past, we have proposed a knowledge base framework called Object-Oriented Knowledge Base (OOKB) which successfully combines the concepts of deductive database, object-oreiented programming and logic programming. In this research, we propose to extend the framework of object-oriented knowledge base with an efficient knowledge base management system. In particular, we propose the following research tasks: Knowledge Management: In our research, we propose the necessary constructs to describe our knowledge (meta-knowledge) about the system-level objects such as integrity constraints, deductive laws, classes, methods, and transactions. We also propose to employ the meta-level knowledge to ensure the quality of knowledge in a dynamic environment where knowledge is updated frequently and to facilitate to control of knowledge bases. Query Processing/Optimization: In the context of query processing, we propose a program transformation approach to avoid exhaustive search. The transformation system employs rewriting rules to substitute pieces of a query with some semantically equivalent sub-queries that can be solved more efficiently. Subsequently, the conjuncts in the transformed query are reorderd dynamically to minimize the evaluation cost. Concurreny Control: Coupled with the meta-knowledge constructs proposed in (a), we propose that concurrency control (among concurrent operations to knowledge as well as data) can be object-oreiented such that it can be customized according to the semantics of objects. Also, we classify the possible conflicts into six categoried and propose that concurrency control should be performed at knowledge level as well as object level.

在过去，我们提出了一个知识库框架， 面向对象的知识库（OOKB）成功地结合了 演绎数据库、面向对象程序设计 逻辑编程。 在这项研究中，我们建议扩大 面向对象的知识库框架， 知识库管理系统。 特别是，我们建议 以下研究任务： 知识管理：在我们的研究中，我们提出了必要的 我们的知识（元知识）， 系统级对象，如完整性约束、演绎 定律、类、方法和交易。 我们亦建议 运用元级知识来确保 在知识不断更新的动态环境中的知识 并促进对知识库的控制。 查询处理/优化：在查询的上下文中 处理，我们提出了一个程序转换的方法，以避免 彻底的搜索 转换系统采用重写 用一些语义上的规则替换查询的片段， 可以更有效地解决的等效子查询。 随后，对转换后的查询中的合取词进行重新排序 动态地最小化评估成本。 并发控制：与元知识结构相结合 在（a）中提出，我们建议并发控制（在 对知识和数据的并发操作）可以被 面向对象，以便可以根据 对象的语义。 此外，我们还对可能的冲突进行了分类， 并提出并发控制应 在知识级和对象级执行。