Research
My research interest lies primarily in the field of theoretical computer science, with a particular focus on the applicability of formal methods to real-world problems.
Bachelor Thesis (Practical Equivalence Checking of P4 Packet Parsers)
Abstract In computer networks, packets function as structured containers for transmitting information. Packet parsers validate these packets in devices like packet switches and firewalls. P4 is a domain-specific programming language that can specify such parsers in software-defined networks. As with any software system, bugs may be present, or improvements are to be made. These subsequent modifications may introduce discrepancies relative to the intended specification. To address this, I introduced Octopus, a tool for the practical equivalence checking of packet parsers written in P4. Here, practical signifies a runtime suitable for execution on a typical, modern personal computer. The tool additionally produces a certificate that enables manual verification of the result’s correctness.
An open-access copy of my thesis is available in the LIACS theses repository.
Publications
-
Octopus: Practical Equivalence Checking of P4 Packet Parsers Open AccessJort van Leenen, Tobias Kappé
Computer Aided Verification (CAV 2026), Part I , Lecture Notes in Computer Science (LNCS), pp. 198–211, Springer, 2026. Lisbon, Portugal, 26–29 July 2026.
Awards
-
1st Place, 9th AI Settlement Generation Challenge (GDMC)IEEE Conference on Games (CoG 2026), Madrid, Spain, September 2026.
Competing as team Merge Conflict together with Robert Salden, Lana van Sprang, Niels Versteeg, and Jasper van der Zwet, our AI-driven Minecraft settlement generator achieved the highest overall score in the competition, including the highest aesthetics score ever awarded in the challenge's history. The project was developed as part of Mike Preuss's Modern Game AI course at LIACS.