Reinforcement Learning with Neural Networks: A Survey

Reinforcement learning (RL) comes from the self-learning theory. RL can autonomously get optional results with the knowledge obtained from various conditions by interacting with dynamic environment. It allows machines and software agents to automatically determine the ideal behavior within a specific context, in order to maximize its performance. Neural network reinforcement learning is most popular algorithm. Advantage of using neural network is that it regulates RL more efficient in real life applications. In this paper, we firstly survey reinforcement learning theory and model. Then we present various main RL algorithms. Then we discuss different neural network RL algorithms. Finally we introduce some application of RL and outline some future research of RL with NN.