Planning and Learning in BDI Agents

BDI Agent 中的规划和学习

• 批准号: LP0560702
• 负责人: Prof Lin Padgham
• 金额: $ 15.62万
• 依托单位: RMIT University
• 依托单位国家: 澳大利亚
• 项目类别: Linkage Projects
• 财政年份: 2005
• 资助国家: 澳大利亚
• 起止时间: 2005-05-31 至 2008-05-30
• 项目状态: 已结题
• 来源: https://dataportal.arc.gov.au/RGS/Web/Grants/LP0560702
• 关键词: Planning; Learning; BDI; Agents

Intelligent agents are becoming increasingly important in a range of complex and dynamic applications. BDI agent systems are extremely successful robust and flexible, in real time applications. However they are not actually able to do any traditional planning. Nor do they incorporate learning. This project will incorporate learning and planning capabilities into a BDI system framework, in a focussed and integrated manner that allows benefits from these approaches where appropriate, but without compromising the real-time capability of existing BDI systems. Learned information and new plans will then be fully incorporated for later use by the standard BDI mechanisms.

智能代理在一系列复杂和动态的应用中变得越来越重要。 BDI 代理系统在实时应用中非常成功、稳健且灵活。然而，他们实际上无法进行任何传统的规划。它们也不包含学习。该项目将以集中和集成的方式将学习和规划功能纳入 BDI 系统框架，以便在适当的情况下从这些方法中受益，但不会损害现有 BDI 系统的实时功能。学到的信息和新计划将被完全纳入标准 BDI 机制以供以后使用。