MSc graduates make industry impact with award-winning operational research projects

Scott Jenkins was awarded the prestigious May Hicks Prize for Best Post-graduate Project, and Tom Murarik was named runner-up.

Winner: 'Optimal Strategy for Grid-Scale Batteries Participating in Power Markets and Frequency Response Services'

Scott Jenkins' project, supervised by Lars Schewe from the School of Mathematics, was conducted in collaboration with Flexitricity.

The project focused on developing a data-driven approach to optimise the bidding strategies for grid-scale batteries participating in both wholesale power markets and frequency response services.

Scott developed explainable probabilistic methods to address the core challenge of market uncertainty. These methods provided clear and interpretable results, allowing Flexitricity to optimize battery revenue and effectively communicate strategies to their customers.

Scott's work had three main elements:

• A novel simulation methodology for generating probabilistic auction outcomes.
• A mixed-integer linear optimisation model to determine the optimal strategy for battery utilisation across the day, considering both frequency response auctions and wholesale power markets.
• A pricing methodology was developed to identify the bid prices that maximise expected revenue for Flexitricity in the frequency response auctions.

Runner up: 'Feature Selection for Customer Churn Classification Using Multi-Objective Evolutionary Algorithms'

Tom Murarik's dissertation, undertaken in partnership with Vodafone, focused on the telecommunications industry problem of customer churn.

Tom developed a novel multi-objective evolutionary algorithm called the Feature Selection Genetic Algorithm (FSGA) which he evaluated on real commercial datasets.

The FSGA:

• Employs multi-objective optimization to enhance decision-making, consistently improving model performance by up to 20% and validating feature importance.
• Combines ten feature selection methods and a mix of genetic algorithm techniques and specialised local search operators.

Tom now works for Optrak - where his dissertation continues to be a source of utility - and is working to publish the FSGA with Kit Searle, his School of Mathematics supervisor.