Author/Authors :
Kargar، Mohammad Hosein نويسنده Department of Computer Engineering, Yazd Branch, Islamic Azad University, Yazd, Iran , , R. Seyedaghaee، Naghi نويسنده Department of Computer Engineering, Aliabad Katoul Branch, Islamic Azad University, Aliabad Katoul, Iran , , Rouhinezhad، Fatemeh نويسنده Department of Computer Engineering, Aliabad Katoul Branch, Islamic Azad University, Aliabad Katoul, Iran ,
Abstract :
Reinforcement learning process is a new research topic. It is a kind of learning techniques. The aim of reinforcement learning problem in robot investigation is, reaching the objectives by receiving the highest reward during the robot moving route (path). This study other than investigating reinforcement learning model in robot movement, presents two other methods for optimal robot movement. In the first method, maximum received reward is considered to select the next movement and in the second method the next movement is selected randomly. Implemental results postulate that the second method has higher efficiency than the first one.
