Please contact our internal team members personally for additional thesis topics or send an email to dnet-teaching(at)rub.de.
Flow prolongation is a technique that can improve the quality of performance bounds derived with deterministic network calculus. During the analysis it is, however, not possible to know the most beneficial prolongation among all alternatives. An exhaustive enumeration was shown not scale well (see literature below).
Literature
- Fabien Geyer, Alexander Scheffler and Steffen Bondorf. Network Calculus with Flow Prolongation - A Feedforward FIFO Analysis enabled by ML. In IEEE Transactions on Computers, Special Issue on Real-time Systems, 2022. [@IEEE, bib, dataset]
- Fabien Geyer, Alexander Scheffler and Steffen Bondorf. Tightening Network Calculus Delay Bounds by Predicting Flow Prolongations in the FIFO Analysis. In Proc. of IEEE RTAS 2021. [@IEEE, bib, dataset]
- Steffen Bondorf. Better Bounds by Worse Assumptions — Improving Network Calculus Accuracy by Adding Pessimism to the Network Model. In Proc. of the IEEE International Conference on Communications (ICC), May 2017. [@IEEE]
Graph Neural Networks (GNNs) can be used to speed up the performance analysis with Deterministic Network Calculus (DNC). There are different topics available in the area of GNNs in DNC that extend the existing literature.
Literature
- Fabien Geyer, Alexander Scheffler and Steffen Bondorf. Network Calculus with Flow Prolongation - A Feedforward FIFO Analysis enabled by ML. In IEEE Transactions on Computers, Special Issue on Real-time Systems, 2022. [@IEEE, bib, dataset]
- Fabien Geyer and Steffen Bondorf. DeepTMA: Predicting Effective Contention Models for Network Calculus using Graph Neural Networks. In Proc. of the 38th IEEE International Conference on Computer Communications (INFOCOM 2019), April 2019. [@IEEE, bib, dataset]
We provide a short (not complete) list of do’s and don’ts that are relevant for writing a succesful thesis. Only accessible with your RUB account GitLab@RUB