A Discount-Based Time-of-Use Electricity Pricing Strategy for Demand Response With Minimum Information Using Reinforcement Learning

Demand Response (DR) programs show great promise for energy saving and load profile flattening. They bring about an opportunity indirect control of end-users’ demand based on different price policies. However, the difficulty in characterizing price-responsive behavior customers is a significant challenge towards optimal selection these This paper proposes Aggregator (DRA) transactive policy generation by combining Reinforcement Learning (RL) technique aggregator side with convex optimization problem customer side. The proposed DRA can maintain users’ privacy exploiting DR as only source information. In addition, it avoid mistakenly penalizing users offering discounts incentive to realize satisfying multi-agent environment. With ensured convergence, resultant capable learning adaptive Time-of-Use (ToU) tariffs generating near-to-optimal Moreover, this study suggests off-line training procedure that deal issues related convergence time RL algorithms. suggested process notably expedite and, turn, enable online applications. developed method applied set residential agents order benefit them regulating their thermal loads according generated efficiency approach thoroughly evaluated from standpoint terms shifting comfort maintenance, respectively. Besides, superior performance selected represented through comparative study. An additional assessment also conducted use coordination algorithm validate competitiveness recommended program. multifaceted evaluation demonstrates designed scheme significantly improve quality aggregated low reduction aggregator’s income.