A Call for Explainability: The Path to Trustworthy AI
The President's Medal recognises our work on the pioneering Dynamic Reserve Setting (DRS) project, developed in partnership with The Electricity System Operator (ESO). By applying advanced operational research (OR) and explainable AI, our model can help control room engineers balance efficiency, security, and cost in our rapidly evolving energy landscape.
This is the second time we have received the President’s Medal. Our first award came in 2018 for our work supporting the Federal Communications Commission (FCC) on the ground-breaking Broadcast Incentive Auction, which reshaped spectrum management in the United States. That project generated nearly $20 billion in revenue and enabled the rollout of next-generation mobile services.
More details about that project can be found below:
Both the FCC auction and the DRS project demonstrate the practical impact of operational research when applied to real-world problems.
While the projects differ in scope—one addressing spectrum management, the other grid stability—both required a deep understanding of complex systems, rigorous analytical methods, and close collaboration with our partners.
Dynamic Reserve Setting: Innovation in Action
The Dynamic Reserve Setting project represents a step forward in energy system management. With net-zero objectives requiring increased usage of renewable energy sources such as wind and solar, the challenge of balancing supply and demand becomes more complex. Renewable sources, while essential to reducing carbon emissions, can be unpredictable, and this variability makes balancing supply and demand on the energy grid more complex.
Working closely with National Grid ESO, we developed a data-driven approach to dynamically set electricity reserves based on real-time conditions. Using operational research techniques and explainable AI, the system allows grid operators to make faster, more informed decisions. By optimising reserve settings, we’ve helped improve efficiency and reduce the operational costs of balancing the grid.
Explainable AI played a critical role in this project, offering transparency into how the model makes its predictions. This level of clarity is crucial for decision-makers who need to trust the system and act on its recommendations. The result is a more flexible and responsive grid management system that not only adapts to changing conditions but does so in a way that is both transparent and efficient.
'Our Dynamic Reserve Setting project leverages the power of Artificial Intelligence to enable dynamic reserve allocation based on real-time data. Within 2 hours of being trialled in our Control Room, DRS informed the decision to cut 1GW of excess reserve, reducing the cost of operating the system and therefore consumer bills . We are so pleased that our work with Smith Institute on DRS has been awarded the Operational Research Society’s President’s Medal.'
