Formalization and Automation of Reasoning about Actions

动作推理的形式化和自动化

• 批准号: 9732744
• 负责人: Vladimir Lifschitz
• 金额: $ 36.9万
• 依托单位: University of Texas at Austin
• 依托单位国家: 美国
• 项目类别: Continuing Grant
• 财政年份: 1998
• 资助国家: 美国
• 起止时间: 1998-09-01 至 2003-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=9732744&HistoricalAwards=false
• 关键词: Formalization; Automation; Reasoning; about; Actions

This project is applying ideas developed in recent research into action languages, on the logic of causal explanation, and on satisfiability planning to the formalization and automation of reasoning tasks in a ide range of action domains. The new formalizations must take into account the possibility of complex interactions between cuncurrent actions, the role of the exact duration of an action, the role of knowledge as a prerequisite for executing an action, and the possibility of creating and destroying objects in the process of executing actions. Some of the action domains being investigated include realistic representations of the physical space in which agents move around and are significantly larger than the typical "toy worlds" that research on actions has concentrated on in the past. The result will be another step toward achieving the goals of a theory of commonsense knowledge and reasoning - the reasoning processes used by humans in everyday life, rather than just the reasoning used in mathematical theorems. http://www.cs.utexas.edu/users/vl/

这个项目是应用在最近的研究思想发展到行动语言，因果解释的逻辑，并在满足性规划的形式化和自动化的推理任务的ide范围内的行动域。 新的形式化必须考虑到当前行动之间复杂的相互作用的可能性，行动的确切持续时间的作用，知识作为执行行动的先决条件的作用，以及在执行行动的过程中创建和破坏对象的可能性。 一些正在调查的行动域包括现实的代表性的物理空间中，代理人四处走动，显着大于典型的“玩具世界”的研究集中在过去的行动。 其结果将是实现常识知识和推理理论目标的又一步-人类在日常生活中使用的推理过程，而不仅仅是数学定理中使用的推理。 http://www.cs.utexas.edu/users/vl/