VidyutVanika21: An Autonomous Intelligent Broker for Smart-grids

Sanjay Chandlekar, Bala Suraj Pedasingu, Easwar Subramanian, Sanjay Bhat, Praveen Paruchuri, Sujit GujarIJCAI 2022 · OpenReview

Abstract

An autonomous broker that liaises between retail customers and power-generating companies (GenCos) is essential for the smart grid ecosystem. The efficiency brought in by such brokers to the smart grid setup can be studied through a well-developed simulation environment. In this paper, we describe the design of one such energy broker called VidyutVanika21 (VV21) and analyze its performance using a simulation platform called PowerTAC (PowerTrading Agent Competition). Specifically, we discuss the retail (VV21–RM) and wholesale market (VV21–WM) modules of VV21 that help the broker achieve high net profits in a competitive setup. Supported by game-theoretic analysis, the VV21–RM designs tariff contracts that a) maintain a balanced portfolio of different types of customers; b) sustain an appropriate level of market share, and c) introduce surcharges on customers to reduce energy usage during peak demand times. The VV21–WM aims to reduce the cost of procurement by following the supply curve of the GenCo to identify its lowest ask for a particular auction which is then used to generate suitable bids. We further demonstrate the efficacy of the retail and wholesale strategies of VV21 in PowerTAC 2021 finals and through several controlled experiments.

Introduction

In the modern smart grid, energy brokers act as critical intermediaries between retail customers (households, offices) and wholesale power generators (GenCos). This role presents a three-fold challenge: navigating retail competition to gain market share without incurring peak-demand penalties, managing wholesale volatility in periodic auctions, and maintaining grid balance to avoid expensive penalties. Previous agents often relied on heuristics that struggled to adapt to varying competitor counts or specific generator behaviors. VidyutVanika21 (VV21) addresses these challenges through a modular, game-theoretic approach.

System Architecture and Methodology

VV21 is structured into two primary modules that handle distinct market environments, supported by a suite of data repositories for weather, wholesale, balancing, and tariff information.

The architecture of VidyutVanika21 showing the interaction between repositories, modules, and the game server.

1. Retail Module (VV21–RM)

Rather than maximizing customer count, VV21–RM optimizes for a "sweet spot" in market share (typically ~38-45% in a 5-player game). This is achieved by modeling the competition as a two-player zero-sum game against a "meta-opponent." This strategy prevents the "monopoly trap," where excessive market share leads to non-linear penalties during peak hours. The module uses a Tariff Enhancer to adjust prices and a Tariff Designer to create Time-of-Use (ToU) contracts that encourage customers to shift usage away from peak periods.

2. Wholesale Module (VV21–WM)

The wholesale strategy centers on the Supply Curve Follower (SCF). By analyzing "uncleared asks" from previous auctions, VV21 estimates the lowest price the primary GenCo is willing to accept. This estimated price serves as a ceiling for strategic bidding: the agent bids low when delivery is far off to capture cheap energy from small sellers, and moves closer to the GenCo's expected price as the deadline nears to ensure demand is met and balancing penalties are avoided.

Experimental Results

VV21 was evaluated in the PowerTAC 2021 finals across 386 games. The agent demonstrated superior performance, winning the tournament by a significant margin. Key metrics include a mean energy procurement price of -$54.87/MW—79% cheaper than the next best broker—and a market-leading income-to-cost ratio of 1.67.

Comparison of financial performance metrics across VV21 and other top-tier brokers (TB1-TB6).

Analysis of the results shows that VV21's success is rooted in its ability to "lose small" in unfavorable scenarios. Ablation studies further confirmed that removing the Retail Module or Wholesale Module reduced profits to approximately 50% and 60% of the full agent's performance, respectively, proving that both strategic components are essential for dominance in the smart-grid ecosystem.

Conclusion and Future Work

VidyutVanika21 proves the efficacy of combining game theory with supply-curve modeling for autonomous energy brokerage. While highly successful, the current wholesale strategy assumes predictable GenCo behavior. Future work will focus on developing more granular models for diverse opponent types and improving the integration of highly variable renewable energy sources into the retail portfolio.

Summary updated

Abstract

An autonomous broker that liaises between retail customers and power-generating companies (GenCos) is essential for the smart grid ecosystem. The efficiency brought in by such brokers to the smart grid setup can be studied through a well-developed simulation environment. In this paper, we describe the design of one such energy broker called VidyutVanika21 (VV21) and analyze its performance using a simulation platform called PowerTAC (PowerTrading Agent Competition). Specifically, we discuss the retail (VV21–RM) and wholesale market (VV21–WM) modules of VV21 that help the broker achieve high net profits in a competitive setup. Supported by game-theoretic analysis, the VV21–RM designs tariff contracts that a) maintain a balanced portfolio of different types of customers; b) sustain an appropriate level of market share, and c) introduce surcharges on customers to reduce energy usage during peak demand times. The VV21–WM aims to reduce the cost of procurement by following the supply curve of the GenCo to identify its lowest ask for a particular auction which is then used to generate suitable bids. We further demonstrate the efficacy of the retail and wholesale strategies of VV21 in PowerTAC 2021 finals and through several controlled experiments.
BibTeX
@inproceedings{chandlekar2022vidyutvanika,
  author = {Sanjay Chandlekar and Bala Suraj Pedasingu and Easwar Subramanian and Sanjay Bhat and Praveen Paruchuri and Sujit Gujar},
  title = {{VidyutVanika21: An Autonomous Intelligent Broker for Smart-grids}},
  booktitle = {International Joint Conference on Artificial Intelligence},
  year = {2022},
  url = {https://openreview.net/forum?id=8VA0rM2vcF},
}
Figure