A Reinforcement Learning System Mimicking Brain Information Processing

在这项研究中，一个模拟大脑信息处理的强化学习系统被开发出来，在一个多智能体系统中出现合作行为。所开发的系统具有其他智能体的状态预测功能，系统结构的自主构造功能，以及适应时变环境的能力。通过计算机仿真和真实机器人实验表明，该算法能够产生协作行为。

In this research, a reinforcement learning system mimicking brain information processing has developed to emerge cooperative behavior in a multi-agent system. The developed system has a state predictive function of other agents, an autonomous construction function of the system structure, and also the ability to adapt a time-varying environment. It was shown that it could emerge cooperative behavior through both computer simulations and real robot experiments.