A fast and scalable algorithm for scheduling large numbers of devices under real-time pricing

Abstract

Real-time pricing (RTP) is a financial incentive mechanism designed to encourage demand response (DR) to reduce peak demand in medium and low voltage distribution networks but also impacting the generation and transmission system. Though RTP is believed to be an effective mechanism, challenges exist in implementing RTP for residential consumers wherein manually responding to a changing price is difficult and uncoordinated responses can lead to undesired peak demand at what are normally off-peak times. Previous research has proposed various algorithms to address these challenges, however, they rarely consider algorithms that manage very large numbers of houses and devices with discrete consumption levels. To optimise conflicting objectives under RTP prices in a fast and highly scalable manner is very challenging. We address these issues by proposing a fast and highly scalable algorithm that optimally schedules devices for large numbers of households in a distributed but non-cooperative manner under RTP. The results show that this algorithm minimises the total cost and discomfort for 10,000 households in a second and has a constant computational complexity.