I am working as a Senior Scientist at the Netherlands Organisation for Applied Scientific Research (TNO). My current work focuses on artificial intelligence and optimization in the area of urban mobility. Before joining TNO I was a PhD candidate within the Algorithmics group at Delft University of Technology.
This webpage contains information about my publications and software. More recent information can be found on Google Scholar and LinkedIn.
Information gathering in POMDPs using active inferenceErwin Walraven, Joris Sijs and Gertjan J. Burghouts |
The Dial-a-Ride problem with meeting points: A problem formulation for shared demand–responsive transitLianne Cortenbach, Konstantinos Gkiotsalitis, Eric van Berkum and Erwin Walraven |
Open-World Visual Reasoning by a Neuro-Symbolic Program of Zero-Shot SymbolsGertjan Burghouts, Fieke Hillerström, Erwin Walraven, Michael van Bekkum, Frank Ruis, Joris Sijs, Jelle van Mil and Judith Dijk |
Parking in Macroscopic Transport Models: Modelling Parking Capacities in Traffic AssignmentDawn Spruijtenburg, Erwin Walraven, Reinier Sterkenburg and Marieke van der Tuin |
PERFEX: Classifier Performance Explanations for Trustworthy AI SystemsErwin Walraven, Ajaya Adhikari and Cor J. Veenman |
Building digital twins of cities using the Inter Model Broker frameworkWalter Lohman, Hans Cornelissen, Jeroen Borst, Ralph Klerkx, Yashar Araghi and Erwin Walraven |
Simultaneous modelling of access, egress & transit line choice for public transportMarieke van der Tuin, Han Zhou and Erwin Walraven |
The Short-Term Potential of Artificial Intelligence for Traffic ManagementHenk Taale, Erwin Walraven, Dawn Spruijtenburg and Isabel Wilmink |
Anomaly Detection in an Open World by a Neuro-symbolic Program on Zero-shot SymbolsGertjan J. Burghouts, Fieke Hillerström, Erwin Walraven, Michael van Bekkum, Frank Ruis and Joris Sijs |
Constrained Multiagent Markov Decision Processes: a Taxonomy of Problems and AlgorithmsFrits de Nijs, Erwin Walraven, Mathijs de Weerdt and Matthijs T. J. Spaan |
Point-Based Value Iteration for Finite-Horizon POMDPsErwin Walraven and Matthijs T. J. Spaan |
Planning under Uncertainty in Constrained and Partially Observable EnvironmentsPhD dissertation, Delft University of Technology, 2019. |
Column Generation Algorithms for Constrained POMDPsErwin Walraven and Matthijs T. J. Spaan |
Bootstrapping LPs in Value Iteration for Multi-Objective and Partially Observable MDPsDiederik M. Roijers, Erwin Walraven and Matthijs T. J. Spaan |
Accelerated Vector Pruning for Optimal POMDP Solvers (supplement)Erwin Walraven and Matthijs T. J. Spaan |
Bounding the Probability of Resource Constraint Violations in Multi-Agent MDPs (supplement)Frits de Nijs, Erwin Walraven, Mathijs M. de Weerdt and Matthijs T. J. Spaan |
Planning Under Uncertainty for Aggregated Electric Vehicle Charging with Renewable Energy SupplyErwin Walraven and Matthijs T. J. Spaan |
Traffic flow optimization: A reinforcement learning approachErwin Walraven, Matthijs T. J. Spaan and Bram Bakker |
Planning under Uncertainty for Aggregated Electric Vehicle Charging using Markov Decision ProcessesErwin Walraven and Matthijs T. J. Spaan |
Planning under Uncertainty with Weighted State ScenariosErwin Walraven and Matthijs T. J. Spaan |
A Scenario State Representation for Scheduling Deferrable Loads under Wind UncertaintyErwin Walraven and Matthijs T. J. Spaan |
Traffic Flow Optimization using Reinforcement Learning (abstract)Proceedings of the 26th Benelux Conference on Artificial Intelligence, pp. 211–212, 2014. |
Traffic Flow Optimization using Reinforcement LearningMaster's thesis, Delft University of Technology, 2014. |
Enhancing SAT Based Planning with Landmark KnowledgeJan Elffers, Dyan Konijnenberg, Erwin Walraven and Matthijs T. J. Spaan |
My accelerated version of incremental pruning for POMDPs is available in SolvePOMDP, which is an open source software toolbox containing exact and approximate algorithms for POMDPs.
The ConstrainedPlanningToolbox contains a collection of algorithms for constrained multi-agent planning under uncertainty.
The finite-horizon POMDP solver FiVI is not available as stand-alone implementation, but the source code is part of the ConstrainedPlanningToolbox. It can be found in this file.
The source code of PERFEX and my implementation of active inference are available on GitHub.