In-Class Exercises

Using reinforcement learning to learn a optimal policy for energy grids that is the most energy efficient and reduces losses due to fluctuations in load. Smart grids are a very important sector of industry which has yet to see the benefits of machine learning primarily due to its antiquity and slow modernization. However, thanks to growth in smart sensor adoption, it is becoming more practical to use reinforcement learning techniques (which is a form of approximate dynamic programming) to solve real and important problems in our smart grid.

Worked with Mohammed Elmzaghi for Smart Grid Senior Design Project.