Filip Rydin
PhD Student · Systems & Control · Chalmers University of Technology

Filip Rydin

filipry@chalmers.se

About

I am a PhD student in the Division of Systems and Control at the Department of Electrical Engineering, Chalmers University of Technology, supervised by prof. Balazs Kulcsar and prof. Morteza Haghir Chehreghani.

My research focuses on developing learning-based methods for combinatorial optimization, with a particular emphasis on vehicle routing problems. I work at the intersection of graph neural networks and reinforcement learning, designing algorithms that can learn to solve complex routing and scheduling tasks more efficiently than traditional heuristics.

More broadly, I'm interested in graph-based learning and optimal control, exploring how modern machine learning techniques can tackle problems that have traditionally been the domain of operations research and classical optimization.

Interested in collaboration? I'm always open to discussing research ideas and potential projects. Feel free to reach out via email at filipry@chalmers.se.

Research Interests

Combinatorial Optimization
Graph Neural Networks
Reinforcement Learning
Vehicle Routing
Optimal Control
Deep Learning
Stochastic Systems

Publications

Conference Papers

F. Rydin, A. Lischka, J. Wu, M. Haghir Chehreghani, B. Kulcsar
Beyond Simple Graphs: Neural Multi-Objective Routing on Multigraphs
14th International Conference on Learning Representations (ICLR), 2026

Workshop Papers

K. Bågmark, F. Rydin
Neural likelihood surrogates for parameter inference via log-density PDE
ICLR Workshop on AI and Partial Differential Equations, 2026

Preprints

F. Rydin, M. Haghir Chehreghani, B. Kulcsar
Two-Stage Learned Decomposition for Scalable Routing on Multigraphs
Preprint, 2026
K. Bågmark, F. Rydin
High-dimensional Bayesian filtering through deep density approximation
Preprint, 2025
A. Lischka, F. Rydin, J. Wu, M. Haghir Chehreghani, B. Kulcsar
A GREAT Architecture for Edge-Based Graph Problems Like TSP
Preprint, 2025
K. Bågmark, A. Andersson, S. Larsson, F. Rydin
A convergent scheme for the Bayesian filtering problem based on the Fokker-Planck equation and deep splitting
Preprint, 2024

Teaching & Supervision

Teaching Activities at Chalmers

  • EEN050 Robust and Nonlinear Control — teaching assistant (2024, 2025, 2026).
  • ESS101 Modelling and Simulation — teaching assistant (2025, 2026).
  • SSY052/SSY310 Automatic Control — teaching assistant (2025, 2026).
  • SSY285 Linear Control System Design — teaching assistant (2025).

Ongoing M.Sc. Thesis Supervision

  • Xinwei Wu — Graph Neural Networks for Traffic Flow Prediction under Network Reconfiguration Proposal (pdf)