多机器人动态编队的强化学习算法研究

在人工智能领域中，强化学习理论由于其自学习性和自适应性的优点而得到了广泛关注．随着分布式人工智能中多智能体理论的不断发展，分布式强化学习算法逐渐成为研究的重点．首先介绍了强化学习的研究状况，然后以多机器人动态编队为研究模型，阐述应用分布式强化学习实现多机器人行为控制的方法．应用SOM神经网络对状态空间进行自主划分，以加快学习速度；应用BP神经网络实现强化学习，以增强系统的泛化能力；并且采用内、外两个强化信号兼顾机器人的个体利益及整体利益．为了明确控制任务，系统使用黑板通信方式进行分层控制．最后由仿真实验证明该方法的有效性．

Research on Dynamic Team Formation of Multi-Robots Reinforcement Learning

In the field of artificial intelligence, the reinforcement learning theory is receiving more and more attention with the advantage of its self learning and self adaptability With the development of the multi agent theory in distributed artificial intelligence, the distributed reinforcement learning is becoming the focus of this research In this paper, the research status of the reinforcement learning algorithm is illustrated first Then the multi robots' dynamic team formation is used as the study model to illuminate the hierarchical behavior control of the robots system with the usage of the reinforcement learning In the algorithm explained here, the SOM neural network is used to partition the state space automatically to speed up the learning rate The BP neural network is adopted to realize the reinforcement learning to strengthen the generalization ability The inside reinforcement signal and outside reinforcement signal are employed to represent the interest of the individual robot and the group robots respectively In order to define the task, the multi layer control and the blackboard communication are used in the system Finally, the simulation results are provided to show the validity of the algorithm