If you are interested in the topics of optimisation under uncertainty, distributed optimisation and / or smart electricity networks, and are looking for an undergraduate student project, then contact me and we can discuss ideas. What follows is an example selection of possible project topics.
Accelerated Optimisation Among Agents
Algorithms like the Alternating Direction Method of Multipliers can be used to distribute the computation of an optimisation problem among multiple self-interested agents, arriving at a coordinated set of actions. However, such algorithms can often suffer from slower convergence where they tend to "circle" around an optimal solution rather than move directly toward it.
This project will investigate promising new techniques for enhancing the ADMM algorithm in order to accelerate convergence towards an optimal solution. In particular, we will explore their suitability for distributed solving of Optimal Power Flow problems.
Scaling Interior Point Methods for Optimal Power Flow
Solving Optimal Power Flow (OPF) for real-time operation becomes increasingly difficult as the number of participants and size of the network grows. One of the largest bottlenecks in interior point methods is the repeated solving of sequence of large sparse linear systems.
This project will explore strategies for exploiting OPF-specific structure of the linear systems at the heart of interior point methods.
ENGN8831: Smart Grids
ENGN8527: Smart Grids
ENGN4524: PV Technologies
Reactive Power Demo