Yutong Nie’s research while affiliated with Zhejiang University and other places

Residential loads have great potential to enhance the efficiency and reliability of electricity systems via demand response (DR) programs. One major challenge in residential DR is how to learn and handle unknown and uncertain customer behaviors. In this paper, we consider the residential DR problem where the load service entity (LSE) aims to select an optimal subset of customers to optimize some DR performance, such as maximizing the expected load reduction with a financial budget or minimizing the expected squared deviation from a target reduction level. To learn the uncertain customer behaviors influenced by various time-varying environmental factors, we formulate the residential DR as a contextual multi-armed bandit (MAB) problem, and develop an online learning and selection (OLS) algorithm based on Thompson sampling to solve it. This algorithm takes the contextual information into consideration and is applicable to complicated DR settings. Numerical simulations are performed to demonstrate the learning effectiveness of the proposed algorithm.

Residential load demands have huge potential to be exploited to enhance the efficiency and reliability of power system operation through demand response (DR) programs. This paper studies the strategies to select the right customers for residential DR from the perspective of load service entities (LSEs). One of the main challenges to implement residential DR is that customer responses to the incentives are uncertain and unknown, which are influenced by various personal and environmental factors. To address this challenge, this paper employs the contextual multi-armed bandit (CMAB) method to model the optimal customer selection problem with uncertainty. Based on the Thompson sampling framework, an online learning and decision-making algorithm is proposed to learn customer behaviors and select appropriate customers for load reduction. This algorithm takes the contextual information into consideration and is applicable to complicated DR settings. Numerical simulations are performed to demonstrate the efficiency and learning effectiveness of the proposed algorithm.