Logic, databases and parallel computation (Information and Computer Science)

逻辑、数据库和并行计算（信息和计算机科学）

• 批准号: 8617344
• 负责人: Paris Kanellakis
• 金额: $ 31.83万
• 依托单位: Brown University
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1987
• 资助国家: 美国
• 起止时间: 1987-05-15 至 1991-04-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=8617344&HistoricalAwards=false
• 关键词: Logic; databases; parallel; computation; Information

The goal of this research is to develop and test efficient, parallel evaluation methods for database and knowledge base query languages. We will use the theory of parallel algorithms to identify and realize the potential parallelism in new, logic-based query languages. For this we will investigate the fine-grained parallelism of primitive symbolic operations such as unification, term matching, congruence closure and special fixpoint operators. We will develop and implement parallel algorithms for recursive logical queries on state-of-the-art multiprocessor systems. We will also examine the use of equational theorem-proving techniques for database manipulation and design. The efficient integration of such techniques with recursive logical queries is an important step in combining functional and logic programming in a structured database context. In recent years, basic query languages, such as the relational calculus, have been augmented by adding new programming features. With these new features, database and knowledge base languages are now moving closer to full programming languages. The optimization and efficient evaluation of these languages is crucial, if one is to combine the high performance of database technology with the power of computing methods from artificial intelligence. New parallel (multiprocessor) machines offer the raw computational power for realizing this goal. But to harness that raw power for database and knowledge base applications, completely new algorithmic techniques - adapted to parallel evaluation - are essential. Such techniques will be developed by this research.

这项研究的目标是开发和测试高效， 数据库和知识库查询的并行评价方法 语言我们将使用并行算法的理论来识别 并在新的基于逻辑的查询中实现潜在的并行性 语言 为此，我们将研究细粒度并行性 基本的符号操作，如统一，术语匹配， 同余闭包和特殊不动点算子。我们将开发 并实现递归逻辑查询的并行算法， 最先进的多处理器系统。我们还将研究使用 数据库操作的等式定理证明技术 和设计. 这些技术的有效结合， 递归逻辑查询是组合 结构化数据库环境中的函数和逻辑编程。 近年来，基本的查询语言，如关系 微积分，已经通过增加新的编程功能来增强。 有了这些新特性，数据库和知识库语言 现在正在接近完整的编程语言。优化 有效地评估这些语言是至关重要的， 将数据库技术的高性能与 人工智能的计算方法。新的并行 （多处理器）机器提供原始的计算能力， 实现这一目标。 但是为了利用数据库的原始力量， 基于知识的应用程序，全新的算法技术- 适应并行评价-是必不可少的。 这种技术将 通过这项研究开发。