Reinforcement learning with artificial microswimmers : vimar