Students should have strong background in programming in Java, as well as strong math skills. Students should be majoring in computer science or a closely related field such as mathematics.
Power systems and energy markets are quickly evolving due to new generation and storage technologies, as wells as the integration of new "smart" technologies to monitor and control distribution and consumption activities. One challenge is how to design better energy markets, including better strategies for agents who trade in the markets. This project seeks to design improved automated intelligent trading strategies for agents participating in energy markets. We are competing in the annual Power TAC competition, which is an international competition in which automated trading agents compete to maximize profits in the realistic simulation of an energy market integrating smart grid elements. There are many challenging research problems in implementing an agent for this competition, including improving price and demand predictions using machine learning, designing new bidding strategies for participating in complex auctions, and designing strategies for issuing and updating tariffs to present to potential consumers.
We are working to find ways to improve the performance of automated trading agents in energy markets, including improvements to both market predictions and trading strategies in multiple types of complex auctions.
We will program novel agent designs using the Java programming language, and may use a variety of techniques in artificial intelligence including machine learning and optimization.