Social Profit Optimization with Demand Response Management in Electricity Market: A Multi-timescale Leader-following Approach

TL;DR

This paper proposes a multi-timescale leader-following approach with demand response management to enhance social profit in electricity markets, balancing individual profit motives with societal benefits.

Contribution

It introduces a novel demand function amelioration strategy and projection-based pricing algorithms to improve social profit without altering UCs' self-interests.

Findings

Proposed algorithms achieve near-optimal social profit.

Multi-timescale leader-following improves market efficiency.

Simulation confirms effectiveness via price of anarchy analysis.

Abstract

In the electricity market, it is quite common that the market participants make "selfish" strategies to harvest the maximum profits for themselves, which may cause the social benefit loss and impair the sustainability of the society in the long term. Regarding this issue, in this work, we will discuss how the social profit can be improved through strategic demand response (DR) management. Specifically, we explore two interaction mechanisms in the market: Nash equilibrium (NE) and Stackelberg equilibrium (SE) among utility companies (UCs) and user-UC interactions, respectively. At the user side, each user determines the optimal energy-purchasing strategy to maximize its own profit. At the UC side, a governmental UC (g-UC) is considered, who aims to optimize the social profit of the market. Meanwhile, normal UCs play games to maximize their own profits. As a result, a basic leader-following problem among the UCs is formulated under the coordination of the independent system operator (ISO). Moreover, by using our proposed demand function amelioration (DFA) strategy, a multi-timescale leader-following problem is formulated. In this case, the maximal market efficiency can be achieved without changing the "selfish instinct" of normal UCs. In addition, by considering the local constraints for the UCs, two projection-based pricing algorithms are proposed for UCs, which can provide approximate optimal solutions for the resulting non-convex social profit optimization problems. The feasibility of the proposed algorithms is verified by using the concept of price of anarchy (PoA) in a multi-UC multi-user market model in the simulation.