Studying the effect of the number of layers of a deep neural network in improving the reward of a reinforcement learning robot

Abstract

The Q learning algorithm in reinforcement learning is one of the algorithms that allows the robot to learn the surrounding environment without the need for prior training samples with the principle of reward and punishment for the robot through interaction with the environment. Increasing the number of hidden layers of the deep neural network used and adjusting some of the higher parameters in it can increase the reward of the robot and thus obtain the best path to achieve the goal.