Exploration of Reinforcement Learning to Play Snake Game

In this research, we explore the hypothesis that Reinforcement Learning applications are not amenable to conduct close analysis. With the combination of Deep Reinforcement Learning and neural network, the snake game agent is trained. The dependences of this game are on python, Keras, sci-kit image, and tensor flow. For the training of the agent in the game, SARSA is used to control the moving capabilities of the snake as per the user's desire. Deep Q-learning is used in the learning module and the learning rate of the SARSA algorithm shows the effective results in the different agent iterations.