About me
Hey there, I am Sanjay Chandlekar, a Ph.D. (in Computer Science) student at the International Institute of Information Technology (IIIT), Hyderabad. I am working with Dr. Sujit Gujar in Machine Learning Lab (MLL). My research interests mainly include deep RL, game theory and system design.
My current work is focused on a project called PowerTAC, which is an abbreviation for Power Trading Agent Competition. PowerTAC is an advanced simulation of real-world smart grids. As a participant in this competition, we have to design an autonomous broker agent that can make strategic decisions like bidding in day-ahead auctions of electricity in the wholesale market, publishing attractive and profitable tariffs in the retail market, and balancing supply and demand of electricity, and various others. We apply game theoretically sound RL techniques to make decisions in the wholesale and retail market and deep learning based prediction models to forecast customer demand. The goal is to earn the highest profit among all the participant brokers in order to win games and, subsequently, the competition.
I completed my B. Tech. from the Institute of Technology, Nirma University in 2018 with a degree in Information Technology. I was awarded a gold medal for being the topper in the Information Technology degree program. During my last semester in B.Tech, I did an internship at TCS Innovation Labs in Hyderabad, where I started learning about Auction Theory and Reinforcement Learning (RL). I implemented a set of bidding strategies motivated by RL techniques to sell electricity in the open market on behalf of an energy power plant with the aim of maximizing profits. After finishing my undergrad, I joined TCS Innovation Labs, Hyderabad as a researcher. I started reading more RL and auction theory. Simultaneously, I analyzed European and Australian power markets and designed bidding strategies for energy procurement for these markets. Similarly, I started working on PowerTAC and spent a few months learning about the simulator. During my time in TCS, we deployed various RL-based strategies in the wholesale (Monte-Carlo, Q-Learning) and retail markets (Q-Learning) of PowerTAC , intending to design an autonomous learning broker for smart-grid systems. Furthermore, we deployed various statistical and neural network based time series prediction techniques (e.g., ARIMA, LSTM etc.) to forecast customer usage in the smart grid. Our broker achieved the 4th rank in the PowerTAC2020 competition, which we transcended by becoming the champions in the PowerTAC2021 and PowerTAC2022 tournament.
Timeline
-
17th April, 2024:
Our paper "Optimizing Prosumer Policies in Periodic Double Auctions Inspired by Equilibrium Analysis" got accepted for publication at IJCAI'24.
-
10th April, 2024:
Our paper summarising our work on smart grids from business point of view got accepted for TCS Research's global competition TACTiCS2.0.
-
29th March, 2024:
Our paper "A Novel Bidding Strategy for PDAs using MCTS in Continuous Action Spaces" got accepted for publication at EMAS'24.
-
1-3 February, 2024:
Attended "Google Research Week 2024" at Google India, Bangalore.
-
4-7 January, 2024:
Got selected for "ACM India PhD Clinic" that happened during CODS-COMAD 2024 in IIIT Bangalore. In the PhD clinic, I had insightful discussions with Prof. Vineeth Balasubramanian and Prof. Janardhan Rao Doppa.
-
25 November, 2023:
Our paper "Multi-armed Bandit Based Tariff Generation Strategy for Multi-agent Smart Grid Systems" got published in "Lecture Notes in Computer Science (volume 14378)" book.
-
17 November, 2023:
Our book chapter "VidyutVanika: AI-Based Autonomous Broker for Smart Grids: From Theory to Practice" got published in "Energy Sustainability through Retail Electricity Markets" book.
-
16 November, 2023:
Presented our work at the Policy, Awareness, Sustainability and Systems (PASS) Workshop 2023.
-
19-25 August, 2023:
Presented our paper in-person at the IJCAI'23 conference held in Macau, S.A.R.
-
29 May-2 June, 2023:
Presented our papers in-person at the AAMAS'23 conference and EMAS'23 workshop held in London, UK.
-
20th April, 2023:
Our paper "A Novel Demand Response Model and Method for Peak Reduction in Smart Grids -- PowerTAC" got accepted for publication at IJCAI'23.
-
27th March, 2023:
Our paper "Multi-armed Bandit Based Tariff Generation Strategy for Multi-Agent Smart Grid Systems" got accepted for publication at EMAS'23.
-
3rd January, 2023:
Our paper "A Novel Demand Response Model and Method for Peak Reduction in Smart Grids -- PowerTAC" got accepted for publication at AAMAS'23.
-
15th December, 2022:
Our autonomous broker agent "VidyutVanika22" became the winner of the PowerTAC tournament second consecutive time.
-
September, 2022:
Submitted book chapter titled "VidyutVanika: AI-Based Autonomous Broker for Smart grids -- From Theory to Practice" for publication in Springer.
-
23-29 July, 2022:
Presented our paper in-person at the IJCAI'22 conference held in Vienna, Austria.
-
9-13 May, 2022:
Presented our paper virtually at the AAMAS'22 conference.
-
21st April, 2022:
Our paper "VidyutVanika21: Autonomous Intelligent Broker for Smart-grids" got accepted for publication at IJCAI'22.
-
12th March, 2022:
Our poster "PowerGame: Role of Artificial Intelligence and Game Theory in Smart-grids" won the best lab poster award in IIIT-H R&D Showcase 2022.
-
19th December, 2021:
Our paper "Multi-unit Double Auctions: Equilibrium Analysis and Bidding Strategy using DDPG in Smart-grids" got accepted for publication at AAMAS'22.
-
8-9 December, 2021:
Presented our autonomous broker agent "VidyutVanika21" in PASS2021 workshop.
-
27th October, 2021:
Our autonomous broker agent "VidyutVanika21" became the winner of the PowerTAC 2021 tournament.
-
10th August, 2020:
Started MS by Research in Computer Science at IIIT Hyderabad.
-
25th July, 2019:
Presented our paper "Vidyutvanika18: An autonomous broker agent for smart grid environment" at PASS2019 workshop held in Cologne, Germany.
-
23rd July, 2018:
Joined TCS Innovation Labs as a Researcher.
