U.S.-Japan Cooperative Research: Applied Constraint Logic Programming

这一奖项将使纽约城市大学布鲁克林学院的Ken McAloon教授和同事们能够在两年内与日本东京科学大学的Fumio Mizoguchi教授合作。他们将探索一个新的逻辑研究领域的发展，他们称之为“应用约束逻辑”。将探索这种类型的逻辑在人工智能、运筹学和软件工程中的应用。本研究的目的是为约束逻辑程序设计系统开发顺序和并行的约束求解器，并最终将约束逻辑的概念与人工智能的概念相结合，产生一个面向金融应用的原型决策支持系统。约束逻辑编程将基于规则的系统与其中问题的组成部分被描述为约束的系统相结合，并通过规则将约束链接在一起来表示整个问题。在某些应用中，约束条件比绝对数量条件更自然地表示问题，对于这些问题，基于约束的系统应该比标准编程语言更有效。在这个项目中，日本小组在算术约束求解器和逻辑编程系统开发方面的工作将与美国小组在开发广义线性约束算法和约束语言软件开发方面的经验相结合。这项研究可能会对我们为人工智能任务编程的方法产生重大影响。

This award will enable Prof. Ken McAloon and co-workers of Brooklyn College, City University of New York, to collaborate with Prof. Fumio Mizoguchi of the Science University of Tokyo, Japan, over a period of two years. They will explore the development of a new field of logic study, which they term "Applied Constraint Logic." The application of this type of logic to artificial intelligence, operations research, and software engineering will be explored. The purpose of this research is to develop sequential and parallel constraint solvers for constraint logic programming systems, and ultimately to combine constraint logic concepts with those of artificial intelligence to produce a prototype decision support system for financial applications. Constraint logic programming combines rule-based systems with systems in which components of a problem are stated as constraints, and the problem as a whole is represented by linking the constraints together by means of rules. In some applications constraint conditions represent problems more naturally than absolute quantitative conditions, and for these problems constraint-based systems should be more efficient than standard programming languages. In this project the Japanese group's work on arithmetic constraint solvers, and on development of logic programming systems, will be combined with the experience of the U.S. group on the development of algorithms for generalized linear constraints and on software development in constraint languages. This research could have great impact on our approach to programming for artificial intelligence tasks.