Jamie Fairbrother is Lecturer in Operational Research Optimisation at the Lancaster University Management School, England
PhD in Statistics and Operational Research, Lancaster University (2011-2016)
MRes in Statistics and Operational Research (Distinction), Lancaster University (2010-2011)
MMath (Hons) in Mathematics with Study in Europe (First Class), University of Warwick (2006-2010)
Data science, decision making under uncertainty, data reduction, optimization
Duration
4 Months, October - December 2024
April - May 2025
Host at the University of Passau
Prof. Dr. Marc Goerigk
Chair for Business Decisions and Data Science
The progressing digitalization of business processes has made a large amount of data available, but so far, many application areas struggle to make the best use of this data. Methodologically, such applications often amount to making a decision under uncertainty, that is, we want to make use of available data to find the best possible decision in the here-and-now. We are particularly interested in methods and models for robust decision problems, whereby the aim is to select a decision which mitigates against the worst possible outcome. Such models are important, for example, in safety critical applications which may arise in engineering problems. However such models are notoriously difficult to solve, and one usually must reduce the amount of input data. Selecting a subset of data presents a trade-off: we must use enough data to accurately capture the uncertainty of the situation, but not so big that we cannot solve the model. In this project we develop new techniques for selecting subsets of data to solve such problems, and which quantify the potential deterioration in quality of solutions we obtain by doing this. The particular novelty of our approach is that, rather than applying a general technique, we leverage the structure of the problem itself to reduce the data more effectively. The framework we develop for doing this is applicable to wide class of problems and application areas.