The ECFEB is in the lead for the design of Power TAC, an ongoing trading agent competition (TAC) that provides a test bed for the design of power markets and trading strategies. Competitors come from, in universities and research centres across the world: Westfälische Hochschule; University of Texas at Austin; Chinese University of Hong Kong; Tel Aviv University; Technical University of Crete; and the University of Edinburgh among tens/hundreds of others.
Power TAC is a form of open, modular research which feeds our investigations into the energy markets of the future. If we’re going to get meaningful answers, then our market model needs to encompass a wide range of behaviours, so every year we add new challenges to the competition, asking competitors to devise new strategies for the simulated power markets in the Power TAC model.
We call the participants in our computer model ‘agents’ or, more specifically, ‘trading agents’. We ask simulated trading agents to consider lots of diverse – sometimes even destructive – strategies so that our simulations uncover the full range of possible market outcomes.
TACs have been highly successful in providing this level of diversity and have been used in supply chain management, travel bookings and ad auctions – negotiations for online advertising space in which advertisers and consumers are represented. TACs bring together top researchers from academia and industry who set their agents against each other in a race for the most successful strategies. These regular competitions spur innovation and uncover strategies and interactions that no single research group could have discovered. Researchers analyse the simulation runs at the end of each competition and publish the successful strategies; this means everyone can improve them for the next competition.
Future energy business is not just about efficient, clean, low-cost, renewable energy sources. Efficient price signals can motivate consumers to consume sustainable energy, and power supply systems need better real-time alignment of energy demand and supply.
The Power TAC competition models the high complexity of energy markets. It allows large-scale experimentation agents to act as retail brokers in a power distribution region. They ‘purchase’ power from a wholesale market and from local sources such as homes and businesses with solar panels. They sell power to local customers and into the wholesale market.
Brokers aim to make a profit through trading energy in the wholesale market and selling to customers via a range of tariffs. In this supply chain, the product is infinitely perishable, and supply and demand must be exactly balanced at all times. This is the challenge the retail brokers must solve. They must deal with customers, consumers, producers and ‘pro-sumers’ as well as other brokers acting as intermediaries aiming to maximise profits. The models of customers represent households, small and large businesses, multi-residential buildings, wind parks, owners of solar panels, owners of electric vehicles, and other comsumers and producers of electricity.
The Faculty of Erasmus Centre for Future Energy Business is proud to announce the 2019 Power TAC Competition as one of the premier competitions for open benchmarking platforms. Power TAC, an ongoing trading agent competition (TAC) that provides a test bed for the design of power markets and trading strategies, will invite their global competitors from universities and research centres to enter the competition. Read more
The Faculty of Erasmus Centre for Future Energy Business is proud to present the 2018 Power TAC Competition Final results.
Congratulations to Power TAC 2018 finalists
1. AgentUDE, Universität Duisburg-Essen / DAWIS, Serkan Özdemir
2. VidyutVanika, IIIT Hyderabad (Machine Learning Lab) & TCS, Susobhan Ghosh
3. CrocodileAgent, University of Zagreb, Demijan Grgic
All final results can be viewed here. Thanks to John Collins, Govert Buijs, and Erik Kemperman! Without them we would not be able to run the competition in such a professional manner. We are looking forward to read great papers on the competition, agent strategies, and business models.