Analysis of Decentralized AIMD-based EV Charging Control

Additive Increase and Multiplicative Decrease (AIMD) control strategy has been studied for quite a long time. It was first introduced in 1989 and has been standardized as the Internet's congestion control algorithm. Its simple structure allows easy implementation and adaptation to the problems where fair and stable allocation of a limited resource among a group of agents/users is needed. From the power engineering perspective, this established method can be utilized for the control of electric vehicle (EV) charging, distributed energy sources (e.g. diesel machines, Photovoltaic panels (PV), and wind turbines), and energy storage units, as well as microgrid management and control. In this paper, we present the AIMD method and lay out a mathematical analysis on its stability and rate of convergence with a specific focus on electric vehicle charging. Later, we discuss necessary modifications needed to make the algorithm operate in a fully decentralized manner while still maintaining fairness. Finally, we demonstrate a case study where the algorithm is tested on EV charging control in the IEEE 37-bus distribution test feeder.