Improving the learning process of deep reinforcement learning agents operating in collective heating environments