Towards comfortable and cost-effective indoor temperature management in smart homes: A deep reinforcement learning method combined with future information

•A deep reinforcement learning method for indoor temperature control is proposed.•Predicted mean vote is adopted to evaluate people’s thermal comfort.•The future information obtained by the deep learning method is adopted as the decision information of the agent.•The performance of the proposed method is verified in different situations.•The average energy cost reduces by 24.29% and 23.63% compared to other traditional control methods.