I'm trying to reproduce the work in the paper Demand Response for Home Energy Management Using Reinforcement Learning and Artificial Neural Network. I want to optimize the power consumption for home appliances. The action space is a different power rating for home appliances. My reward function is = -(power rating *electricity price).
I have trained an RL agent using DQN algorithm on Matlab. I have action space that the agent should select from, but my agent always takes the same action irrespective of state. I have checked my reward function and the algorithm does not select the action with the highest reward. Anyone can think of why is the agent behaving this way?
What I'm getting while training:
And my agent always takes the same power rating regardless of the state (electricity price). Why?