In this paper, we present our approach to solve a physics-based reinforcement learning challenge "Learning to Run'' with objective to train physiologically-based human model to navigate a complex obstacle course as quickly as possible.The environment …