My bachelor's project

During my bachelor's project, I worked under the guidance of Professor Enrico Gerding. My project was titled "Energy Storage in the Smart Grid: a Multi-Agent Deep Reinforcement Learning Approach." I presented a novel energy storage system controlled by a reinforcement learning agent for households in a smart grid. The proposed system aimed to optimize electricity trading in a variable tariff environment. The system has been shown through simulations and evaluations to generate significant consumer savings, on average 20.91% of yearly electricity bills, without requiring changes in consumption habits. Furthermore, when integrated with solar panels, the system presented the potential for even greater cost reductions. I have further investigated a Multi-Agent System simulation to analyze between-agents interactions and identify beneficial price-demand relationships. The findings highlight the positive impact of storage on the energy market and demonstrate the advantages for both consumers and network operators. Deep Q Learning was identified as the most effective algorithm, as it adeptly manages the high-dimensional, nonstationary, and stochastic aspects of the problem, bypassing the need for conventional abstract modelling and deterministic rules. My study examined the effects of different storage sizes and agent complexity levels, providing valuable insights into the potential of the proposed solution and its benefits for the wider community.

I received an outstanding grade for this project along with excellent feedback from my supervisor. Some comments are provided below:

  • "In terms of progress, this is one of my best students. Deadlines were kept and the project finished early ..."
  • "The project challenge done at a level of a PhD student ..."
  • "The work completed is novel and could be publishable at a workshop ..."
  • "Pawel clearly shows an excellent understanding of the work, in terms of the technical depth, but also understanding the strengths and weaknesses ..."
  • "Overall this was a successful project which surpassed my expectations."

Encouraged by this feedback, I decided to address the shortcomings in my work and aimed to publish it. I presented my paper at the MLinPL 2023 Conference in a 15-minute talk and poster session, where it received positive feedback. The work was also presented at ICEEEP 2024 in Lille and published in the Springer Trends in Clean Energy Research book.