人工情感在Agent行为选择策略中的应用

研究了人工情感在智能系统中是否可以发挥重要作用．以一生活在虚拟环境中的agent作为研究对象，提出一种基于模糊逻辑的情感模型，并将情感的评估作用与基于CMAC神经网络的联想学习机制相结合，将情感对环境和agent自身状态的评估的变化作为再励信号，用于引导agent的行为选择策略的学习．计算机仿真研究的结果表明，运用该方法可以取得良好的效果．

Study on Artificial Emotion and Its Application in the Behavior Selection Strategy of Agents

In recent years, the fact that emotions are essential in human intelligence has been noticed. This suggests that emotions may play an important role in intelligent systems. The paper presents a fuzzy emotion model for an agent living in a virtual environment and combines the emotions with an associative learning strategy based on CMAC neural network to form the decision-making and self-learning system of the agent, in which the emotions provide reinforcements to direct the learning process. Computer simulation results show that the proposed method can improve the performance of the agent.