Demand modelling in electricity market with day-ahead dynamic pricing

In this paper, we consider a retail electricity market, where the day-ahead dynamic pricing is used and two-way communication is applied. Our objective is to build a demand model that is able to help an electricity retailer understand the customers´ behaviour of using electricity with imperfect information and predict the demand of electricity in the future as accurately as possible. To achieve this objective, we establish the consistent conditions that the demand models must be satisfied and then learn the demand models to estimate the customers´ consumption reaction functions to the retailer´s prices from the available historical data. Simulation results confirm that the proposed demand models can generate a highly accurate prediction of electricity demand.