A Truthful and Near-Optimal Mechanism for Colocation Emergency Demand Response

Abstract

Demand response (DR) has been widely adopted as a strategic plan of the electricity market in maintaining power grid reliability, sustainability, and stability. In a typical emergency DR (EDR) that arises in colocation data centers, participating tenants can reduce their power consumption when the supply of electricity is a shortage and be rewarded with financial compensation. In this paper, we study a mechanism design problem of motivating tenants for colocation EDR (MEDR). To solve the MEDR problem, we present a truthful Fully Polynomial-Time Approximation Scheme (FPTAS) which is theoretically proved deterministic, truthful and near-optimal, and can be approximated within 1 + ϵ for any given ϵ > 0, while the running time is in the polynomial of the number of tenants n and ε. To speed up the calculation of the payments, we further study the Vickrey-Clarke-Groves (VCG) based mechanism. Moreover, we build a MEDR auction system (MEDRAS) and implement all mechanism algorithms for a colocation data center. Comprehensive and detailed experiments have been implemented to validate the efficiency of our proposed mechanisms.