Postdoc in Agent-based Computer Simulation for Future Mobility Systems - DTU Management

You will be affiliated with the Intelligent Transport Systems section at the Department of Technology, Management and Economics (DTU Management) of the Technical University of Denmark. The position is part of a larger EU project entitled “FEDORA - Federation of network optimisation services, simulation foresights, and data alchemy for adaptable, agile, secure, and resilient multimodal traffic management”, funded by the EU Commission’s Horizon Europe Framework.

The successful candidate will work on two integrated research agendas:

• Building an enhance a state-of-the-art agent-based simulation model for assessing future mobility technologies in the Greater Copenhagen region.
• Explore the development of machine-learning based scenario discovery for future mobility policy design.

The simulation model is based on SimMobility open-source framework and creating a new version of the framework with extensions for selected future mobility systems and applying it to the urban Danish case is central to this position. SimMobility is based on activity-based mobility modelling theory, simulating agent-level behavior such as route, departure-time, and mode choice within an activity-based framework. The extensions will allow the evaluation of high-impact interventions including dynamic road pricing and Mobility-as-a-Service (MaaS) solutions.

FEDORA is also advancing the state-of-the-art in machine learning for transport simulation. A core innovation involves Bayesian metamodeling techniques to construct fast surrogate models of the simulation space, enabling efficient scenario analysis without incurring prohibitive computational costs. Unlike standard sequential sampling methods, the project will explore batch-based active learning, prioritizing diversity and informativeness in data selection.

The Postdoctoral position is aimed to start between 1 October 2025 with two year length. The successful candidates will work under the supervision of Associate Professors Carlos Azevedo and Ravi Seshadri. The Intelligent Transport Systems Section belongs to the Transportation Science division of the Department of Technology, Management and Economics (DTU Management) at DTU. The division conducts research and teaching in the field of traffic and transport behavior and planning, with particular focus on behavior modelling, machine learning and simulation.

Responsibilities and qualifications
Your overall aim will be to strengthen the department’s competences within the application of advanced methods for activity-based simulation models and machine learning methods in transportation. You will work in close collaboration with colleagues, and with academic partners in Europe. Your primary tasks will involve formulating, estimating, and testing demand / choice models, implement them in a complex agent-based simulation framework and apply and calibrate them for the Greater Copenhagen Area.

• You will develop, apply and test agent-based mobility simulation frameworks for assessing new mobility solutions and policy impacts. 
• You will design and implement machine-learning algorithms that interact with your simulation framework for scenario discovery, building surrogate models of simulation outputs using probabilistic methods.
• You will collaborate with domain experts across transport modeling, machine-learning, and policy design to ensure scientific and practical relevance. 
• You will contribute to the open-source codebase and documentation for future reuse by the research and policy community.
• You will publish the new findings in the international peer reviewed literature.
• You will present your research at project consortium meetings and international conferences.
• You will communicate effectively with colleagues and be able to work as part of a team to achieve ambitious goals.
• You will be responsible for project milestones and deliverables.
• Teach and supervise MSc student projects

For this position, experience with simulation development and coding is essential. We expect the candidate to have:

• PhD in computer science, machine learning, operations research, transportation engineering or a related field.
• Programming skills in C/C++ and Python, along with experience working with simulation frameworks.
• Strong knowledge of probabilities and statistics.
• Ability to work in a UNIX environment.
• Demonstrated the ability to publish in the international peer-reviewed research literature
• Proven ability to work on interdisciplinary projects and communicate technical concepts