Towards relational hierarchical reinforcement learning in computer games