Demand Response Scheduling Considering Pollutant Diffusion Uncertainty of Industrial Customers

Abstract

The demand response scheduling scheme requires the consideration of both the industrial customers' economic cost and the environmental influences from pollutants. However, the diffusion process of the latter, although of paramount importance, is typically ignored in the existing literature. To address this issue, we propose a demand response scheduling scheme that not only precisely simulates the diffusion process through a spatio-temporal diffusion model, but incorporates the uncertainty into the diffusion trajectories via a Markov decision process. This enables the schedule-maker optimally select the industrial customers to participate in the demand response with a minimum cost while reducing the environmental influences of the pollutants simultaneously. Using it, a deep reinforcement learning approach is further advocated in the optimization procedure to improve the scalability of the proposed method. Simulation results on the modified IEEE-118 test system reveal the validity of the proposed method.