Abstract: This study explores the development of nonlinear activation functions in reinforcement learning through evolutionary computation. Traditional activation functions like Rectified Linear-Unit ...