This thesis is testing out the group of experts regime in the context of reinforcement learning with the aim of reducing the search space used in reinforcement learning. Having tested different abstracion levels with this ...