摘要
设计运行在动态环境中的BDI Agent时,为它设计一个合适的意图再考虑策略是一个关键问题。进行了当目标较多时,如何利用目标的分布情况来指导Agent的意图再考虑的研究,提出了一定区域内群体目标的整体效用的概念,提出根据整体效用选择区域并对Agent的行为进行区域约束的方法。在TILEWORLD环境下用实验来考察基于该方法的几种策略,给出了一种较优的意图再考虑的策略。
One of the key problems in the design of BDI Agent that is situated in dynamic environments is that of finding an appropriate policy for intention reconsideration. When there are lots of goals in the environment, the goals＇ distribution should be useful information for the Agent.This paper investigates how to use the goals＇ distribution to help the Agent to reconsider its intention.A method is presented that enable Agent to calculate the multi-goals＇ utility in certain area and then to use the chosen area to restrict the Agent＇s behavior in future.And then assessed and analyzed several policies based on this method by experiments in the TILEWORLD domain.
出处
《计算机工程与应用》
CSCD
北大核心
2007年第18期90-93,共4页
Computer Engineering and Applications
