Customized Retail Pricing Scheme Design with a Hybrid Data-driven Method

Abstract

<p>Rapid growth of smart metering data in smart grids provides great opportunities for the retailer to design customized price schemes and demand side management (DSM) programs for different customer groups. This paper proposes a hybrid data-driven method of clustering customers' daily load profiles and optimizing different electricity retail plan recommendations for electricity retailers. By combing the user-side information with the risk-aware decision-making framework, specifically using conditional value-at-risk (CVaR) modeling method, the retailer could guarantee its accumulated revenue without doing any harm to the customers' benefit, while guiding their energy consumption behavior instead. Through large-scale experiments, it is observed that a slight increase in the customers' possible payment would be compensated by their big gain in more demand response opportunities. The retailers' profit could also be increased by roughly 49%-51% and 33%-38% with or without enabling demand response programs.</p>