Projects with allocated PhD studentships
Algorithms and Data Analysis
Supervisors: Dr Dimitrios Letsios
One PhD studentship will be allocated to the following project which lies at the intersection of
algorithms, computational optimization, and data science.
Models and Algorithms for Resource Allocation Problems with Machine Learning Predictions
This project aims to design and analyze optimization models and algorithms for temporal resource
allocation problems, e.g. electric power distribution, logistics, and production scheduling problems,
arising in different application domains, including the energy sector, manufacturing and process
engineering [Letsios et al. 2020]. The goal is to effectively assign resources, e.g. machine time and
energy, to activities, so as to optimize performance. Solving instances of such problems may result in
substantial economic benefits. Typically, future resource requirements and customer demand are not
precisely known in advance, but can be predicted using data science and machine learning capabilities
[Bertsimas et al. 2018]. However, these predictions are subject to errors. In this context, determining
efficient algorithms for supporting and automating the resource allocation process is a challenge. To
this end, prior work develops efficient algorithms and optimization models accounting for the time-
varying nature and uncertainty of temporal resource allocation problems [Antoniadis et al. 2020,
Letsios et al. 2021,Manish et al. 2018].
This project aims to (i) develop novel discrete optimization methods for temporal resource allocation
methods and analyze their performance theoretically, (ii) suggest ways to mitigate the effect of
prediction errors in the quality of the obtained solutions, and (iii) evaluate the performance of the
proposed approaches numerically using real data. Prior experience on discrete optimization,
approximation/online algorithms and/or integer programming will be useful.
• [Antoniadis et al. 2020] Antonios Antoniadis, Christian Coester, Marek Eliás, Adam Polak
and Bertrand Simon. Online Metric Algorithms with Untrusted Predictions. International
Conference on Machine Learning (ICML), 2020.
• [Bertsimas et al. 2018] Dimitris Bertsimas, Vishal Gupta, Nathan Kallus. Data-Driven Robust
Optimization. Mathematical Programming, p. 235-292, 2018.
• [Letsios et al. 2020] Dimitrios Letsios, Radu Baltean-Lugojan, Francesco Ceccon, Miten
Mistry, Johannes Wiebe, Ruth Misener. Approximation Algorithms for Process Systems
Engineering. Computers and Chemical Engineering 132, 2020.
• [Letsios et al. 2021] Dimitrios Letsios, Miten Mistry, Ruth Misener. Exact Lexicographic
Scheduling and Approximate Rescheduling. European Journal of Operational Research, 2021.
• [Manish et al. 2018] Manish Purohit and Zoya Svitkina and Ravi Kumar. Improving Online
Algorithms via ML Predictions. Advances in Neural Information Processing Systems
(NeurIPS), p. 9661--9670, 2018.
Data architectures with humans-in-the-loop
Supervisor: Professor Elena Simperl
Fundamental data-centric tasks such as conceptual modelling, content labelling, entity extraction and
query processing are routinely realised as hybrid processes, which consist of human and algorithmic
elements. Examples include any AI system that depends on large amounts of labelled data, interactive
machine learning systems, but also knowledge graphs such as Yago, Wikidata, or DBpedia, which are
created by people alongside a range of more or less sophisticated bots.