arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1303.0795cs.LO
收藏

走向可更新的战略逻辑

Towards an Updatable Strategy Logic

Christophe Chareton, Julien Brunel, David Chemouil

摘要:

本文介绍的是时态多智能体逻辑。这些形式主义中的几个已经提出(ATL-ATL*,ATLsc,SL)。它们能够表达系统中代理的能力,以确保时间属性的满足。特别是,SL和ATLSC使几个代理在一个上下文中混合不同的策略,他们在一个语义游戏中发挥相互作用。我们概括了这种可能性,提出了一种新的形式主义,更新策略逻辑(USL)。在USL中,智能体也可以改进自己的策略。表达能力的提高提高了代理人“可持续能力”的概念。USL是由SL构建的。它主要给SL带来了以下两个方面的修改:语义上,给定状态的后继者不是由每个Agent的一个选择的数据唯一确定的。从语法上讲,我们在语言中引入了一个操作符,称为“unbinder”,它明确地删除了策略与代理的绑定。我们表明,USL是严格的表达比SL。

英文摘要:

This article is about temporal multi-agent logics. Several of these formalisms have been already presented (ATL-ATL*, ATLsc, SL). They enable to express the capacities of agents in a system to ensure the satisfaction of temporal properties. Particularly, SL and ATLsc enable several agents to interact in a context mixing the different strategies they play in a semantical game. We generalize this possibility by proposing a new formalism, Updating Strategy Logic (USL). In USL, an agent can also refine its own strategy. The gain in expressive power rises the notion of "sustainable capacities" for agents. USL is built from SL. It mainly brings to SL the two following modifications: semantically, the successor of a given state is not uniquely determined by the data of one choice from each agent. Syntactically, we introduce in the language an operator, called an "unbinder", which explicitely deletes the binding of a strategy to an agent. We show that USL is strictly more expressive than SL.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑