Wil je wat meer over me te weten komen, maar liever in het Nederlands? Luister naar deze podcast over kunstmatige intelligentie en de energiesector (29 oktober 2020).
I am an Associate Professor on Algorithms for Planning and Optimization and section head of the Algorithmics Group of the TU Delft, Faculty of Electrical Engineering, Mathematics and Computer Science (EEMCS), Department of Software and Computer Technology.
My aim is to develop new algorithmic techniques that cope with challenges prohibiting successful use of algorithmic and artificial intelligence planning and optimization techniques in practice. The most prominent fundamental challenges regarding the planning problems I am working on are:
- Robustness and uncertainty: we often have access to historical data and increasingly accurate predictions; still the future is uncertain, and the aim of planning algorithms is to suggest next steps/decisions that are not only good in expectation, but also allow for changes later, in case of less likely scenarios. In this line of work we build upon the fields of Stochastic Programming, Reinforcement Learning, and AI Planning (e.g. plan repair algorithms).
- Scalability: considering possible future decisions and effects of these increases the theoretical complexity of these already NP-hard problems. State-of-the-art algorithms have runtimes that are exponential in the number of actions/decisions, the horizon, etc., which makes it impossible to use them in practical settings. Furthermore, sometimes there are non-linear relations further complicating the planning and optimization problem. Here we extend and combine techniques from the fields of Operations Research (e.g. Benders decomposition, column generation), Constraint Programming, Decision Diagrams, Evolutionary Algorithms, Surrogate Modelling and Machine Learning.
- Multiple parties: although this is the main approach taken in Operations Research, in practice decision-making is never done in isolation by a single entity with a clear single objective: preferences and information from multiple parties are relevant for the decisions and therefore decisions should not only by cost-efficient but also fair and taken such that self-interested parties have little opportunities to manipulate them. Another related perspective is that all the time many parties make decisions (distributedly) that influence each other and together determine the efficiency/fairness of the system or society. To make progress here we study and extend fundamental research from the fields of Algorithmic Mechanism Design, Social Choice and Multiagent Systems (e.g. multi-agent path finding).
I really enjoy working on these challenges in the context of other science or engineering fields, evaluating my ideas in practice or close-to-real-world simulations. The benefits of confronting state-of-the-art fundamental algorithms with realistic problem models (often together with colleagues from other fields or the industry) is three-fold:
- it provides inspiration, context and motivation for fundamental open problems
- it helps me to acquire relevant data and provide realistic benchmarks to the broader scientific community
- by doing this I can make an impact in practice
Mainly I work on such algorithmic challenges related to the energy transition and in efficient transportation, but also on planning and scheduling problems regarding satellite scheduling and harvesting and logistics of soft fruit chains.
- For example, moving towards sustainable energy, the production costs of electricity become dependent upon the weather. Users planning electricity-consuming activities have individual objectives that are different from those of electricity generators and network operators. As a society we are interested in a pricing scheme for electricity with an equilibrium for electricity consumption and production that optimizes the social welfare. Furthermore, we need algorithms to schedule flexible loads in this uncertain context. To facilitate the development of such algorithms, we have developed an open-source benchmark/simulator called B-FELSA. Please refer to the respective publication on Benchmarking Flexible Electric Loads Scheduling Algorithms.
- In the context of transportation, I'm working closely with the Dutch railways (Nederlandse Spoorwegen) on algorithms for train unit shunting and scheduling of servicing. An important objective for me here is to invite the scientific community to contribute to solving this complex logistic problem by setting up a scientific competition. I'm especially interested in benchmarking different algorithmic techniques (developed by different communities) such as multi-agent path finding algorithms, constraint programming, reinforcement learning, and mathematical programming. Research into this complex logistical problem of train shunting and servicing is made possible by our open-source simulator called TORS. This has been first published in a paper in the demo track at AAMAS.
Within the Algorithmics group we bring together researchers with specific algorithmic and artificial intelligence expertise. We support each other in project acquisition, and complement each other's expertise in executing projects and supervising phd students.
Within the TU Delft organisation, I aim to more broadly facilitate the use of state-of-the-art algorithmic and AI techniques in practice, and in other sciences, please see the AIDU website for the current status of our efforts.
If you're interested in joining me in my quests, please drop me an email.
Mathijs de Weerdt is an associate professor at the Delft University of Technology. After his PhD he received the prestigious VENI grant from NWO to support his research into coordinated planning from 2005 to 2008. He has been a visiting researcher at the Dutch Center for Mathematics and Computer Science (CWI) from 2005 to 2016, Cork in 2006, Southampton between 2012 and 2015 and Duke in 2017. Mathijs currently is scientific advisor at the Dutch national railways (NS), and one of the founders of the TU Delft Rail Institute (2020). He is (co-)promotor of in total 12 PhD students of which 6 have completed their PhD thesis and has been chair of the Algorithmics group since 2018. An important challenge for him is to identify how Artificial Intelligence can contribute to the energy transition and sustainability of our society, and what is needed to speed-up such developments. This is a main driver of his research, but also an important topic in the convergence discussions with Rotterdam and Leiden universities. He is chairman of the working group on Energy & Sustainability of the Dutch AI coalition (NLAIC).
- 1998-2003 PhD Student at Delft University of Technology on Algorithms for Plan Merging in Multi-Agent Systems
- 1994-1998 BSc and MSc., Computer Science, Utrecht University, Intelligent Systems/Algorithmics (cum laude)
Grants as main applicant
- 2020 RVO & TKI TU: Een optimale zacht-fruitketen begint bij de oogst
- 2016 NWO-URSES+: Future-proof Charging
- 2016 NWO ESI-pose: Flexibility in Industry
- 2014 NWO-URSES: Gaming beyond the copper plate
- 2009 Basic-NGI: Dynamic contracting in infrastructures
- 2004 NWO/STW-VENI national personal grant
Grants as co-applicant
- 2021 AI4Bio ICAI Lab with DSM
- 2019 RVO Urban Energy: Flex-DC
- 2019 RVO Urban Energy: Flex met warmte
- 2017 NWO Big data: Real-time data-driven maintenance logistics
- 2016 ERANET Smartgrid Plus: DCSMART
- 2014 AMS Stimulus: From Needs to Knowledge, challenges and opportunities for city-scale crowd sensing in intelligent cities (small grant)
- 2012 TopSector Energy: Smart Energy Systems, WarmteWeb, with Eneco
- 2011 IJCAI Extended Lab Visit, Jim Boerkoel, University of Michigan, Multi-agent temporal planning (small grant)
- 2011 Faculty of EEMCS, internal, SHINE, Sensing Heterogeneous Information Network Environment
- 2010 Marie Curie (as host): Planning under uncertainty for real-world multiagent systems
Prizes and awards
- 2018 ICT.Open poster prize
- 2016 Erasmus Energy Forum Science Award
- 2015 Best teacher award in Computer Science (Delft)
- 2014 Honorable mention Best Dissertation on Algorithms for Simple Temporal Reasoning by dr. Planken, International Conference on Automated Planning and Scheduling (ICAPS), as co-promotor and daily supervisor.
- 2011 Honorable mention ICAPS best paper on Computing All-Pairs Shortest Paths by Leveraging Low Treewidth
- 2011 Nomination for Best Paper, IEEE International Conference on Networking, Sensing and Control (ICNSC).
- 2009 Nomination for Best Paper, Pacific Rim International Conference on Multi- Agents (PRIMA).
- Chair of the ICAPS conference 2018
- Co-chair of conference on CPAIOR 2018
- Organizing Committee (scholarship chair) of the International Joint Conference on Autonomous Agents and Multiagent Systems (AAMAS) in 2014
- (S)PC member of the AAMAS conference 2008-now (SPC in 2009,2010,2012,2016-)
- (S)PC member of the ICAPS conference 2010-now
- (S)PC member of the IJCAI conference 2009-now
- (S)PC member of the AAAI conference 2016-now
- PC member of several other conferences and workshops
- Organizer of the multiagent planning workshop (at AAMAS'07 and ICAPS'08)
- Tutor for the EASSS summer school 2005-2008,2014 (e.g. see the EASSS'14 website for slides and Extra Material on Dynamic and Online Mechanism Design).
- Research fellow of the Research School for Transport, Infrastructures, and Logistics (TRAIL)
- Research fellow of the School for Information and Knowledge Systems (SIKS)
- Board member of the Dutch Association for Theoretical Computer Science (NVTI)
- Board member of the AgTech institute in its inaugural year (2020)
- Coordinator of the lunch lectures for the PowerWeb institute
- Reviewer for many journals, such as Artificial Intelligence (AI), Artificial Intelligence Review (AIRE), Applied Intelligence (APIN), Journal of Autonomous Agents and Multi-Agent Systems (JAAMAS), Annals of Mathematics and Artificial Intelligence (AMAI), IEEE Transactions on SMC, IEEE Transactions on Sustainable Computing, IEEE on Cybernetics, Journal of Scheduling (JOSH), Mathematical Social Sciences (MSS), Multiagent and Grid Systems: An International Journal (MAGS), Transactions on Internet Technology (TOIT), Transportation Research: Part C (TRC), Transactions on Economics and Computation (TEAC), Transactions on Intelligent Systems and Technology (TIST), Theoretical Computer Science (TCS), Journal of Computational Science, Journal of Artificial Intelligence Research (JAIR).
- Grant reviewer and/or grant review committee member for WT-Flanders, Israel Science Foundation, Natural Science and Engineering Research Council of Canada, Czech Science Foundation, Engineering and Physical Sciences Research Council (UK), FP7-STREP (EU) and NWO Klein.
Valorisation and outreach
- 2021 (February): hour long interview (excluding several music intermezzos) for Dutch national radio (NPO Radio 1) BNN VARA, program Gaan! where I answer all kinds of questions from listeners on algorithms in our society (in Dutch; click here for spotify podcast)
- 2020 (November): 10 minute interview for Dutch local radio Haarlem 105 on Algorithms and how to might influence us (in Dutch)
- 2020 (October): 30 minute interview for podcast Snoek op Zolder on AI and the energy system (in Dutch)
- 2019-: Participate in discussions with and within the Dutch AI Coalition NLAIC (chair of working group Energy & Sustainability since December 2020)
- 2019-2020: Algoritmisch Advies Mobiliteit, BIOS groep/Sardina Management: I’m consulting on algorithmic questions regarding mobility such as on algorithms for routing of taxis and for providing personal travel advice.
- 2018 Flexibility in Distribution Networks: Together with Laurens de Vries I wrote an opinion paper for the Energeia newsletter, organized brainstorm session in Delft, and a panel session in an Urban Energy conference to discuss the alternative market concepts and network tariffs for using flexibility in distribution networks to prevent congestion. See my blog entry about this.
- 2018, October 6: interview for De Volkskrant (Sir Edmund) "Eerlijk verdelen we alle stroom"
- 2017–present Scientific Adviser, Dutch National Railways (NS): I’m consulting on algorithmic questions, mostly regarding logistics.
- 2017–2018 Klankbordgroep Dynamo Flexmarktontwikkeling, Alliander. I was member of an advisory board on pilots with a market for flexible electricity use when network capacity limits are reached. We were meeting four times a year for two hours.
- 2017 Stochastic Optimization Model, Jedlix. As part of our joint project (URSES+) on Future-Proof Flexible Charging, we modeled the decision problem of bidding flexibility in charging electric vehicles in the available electricity markets
- 2012–2013 Algorithm for Multi-Modal (Taxi–Train) Rides, Transvision: Both for the proposal for their Valys tender and after having won the tender, I have advised Transvision on several aspects related to optimization (about 40 hours). The algorithm is now used in their daily operations.
- 2012–2013 Serious Game for Multi-Party Decision Making in Road Maintenance. My PhD student Joris Scharpff built a serious game with which several groups from industry and Rijkswaterstaat were trained to increase their awareness regarding the importance of coordinating road maintenance activities. See also his article about a serious gaming experiment on road maintenance planning.
I aim to make code and data public for reproducibility and to support follow-up work. Please see here for the software and data reported in the publications of our group.
- Just started with a project named An optimal soft-fruit chain starts with the harvest (2020-2024)
- Promotor of Ksenija Stepanovic in the Flex-Heat project (2020-2024)
- Promotor of Grigorii Veviurko in the DC-Flex project (2020-2024)
- Promotor of Jesse Mulderij in a project with NS (2019-2024)
- Promotor of Grigory Neustroev in the NWO FlexI project (2017-2021)
- Promotor of Longjian Piao in the EU DCSMART project (2016-2020)
- Supervisor of Laurens Bliek as a post-doc in the project on Real-time data-driven maintenance logistics (2018-2021), see this blog post.
- Supervisor of HE Lei, a PhD student shared with National University of Defense Technology to coordinate and schedule multiple agile autonomous satellites in a constellation (two years in Delft on a CSC scholarship 2017-2019)
- Project leader of the NWO URSES+ Future-Proof Flexible Charging project involving one post-doc position (German Morales, later Natalia Romero) and a programmer (Koos van der Linden) (2017-2019)
- Project leader and promotor of Rens Philipsen in the NWO USES GCP project (tbd)
- Co-promotor of Mengxiao Wu and Jasper Hoogland from CWI, group on Intelligent Autonomous Systems (tbd)
- Project leader and promotor of Joris Scharpff in the NGI project on Dynamic Contracting in Infrastructures (defence November 20, 2020)
- Project leader and promotor of Frits de Nijs in the project on Dynamic Capacity Control and Balancing in the Medium Voltage Grid with Alliander (defence 2019) on Resource-Constrained Multi-agent Markov Decision Processes
- Co-promotor of Jaume Jordan at Universitat Politècnica de València (defence August 2017) on Non-Cooperative Games for Self-Interested Planning Agents (cum laude)
- Co-promotor of Gleb Polevoy in the SHINE project (defence 2016) on Participation and Interaction in Projects
- Co-promotor of Leon Planken (defence 2013) on Algorithms for Simple Temporal Reasoning
- Co-promotor of Tamas Mahr (defence 2011) on Vehicle Routing under Uncertainty
- Co-promotor of Sicco Verwer (defence 2010) on Efficient Identification of Timed Automata: Theory and Practice
I did a study when visiting Southampton for about 10 weeks. Among other things, this resulted in a journal publication on routing and en-route charging of electrical vehicles. See this fun video on Minimising Queues at Electric Vehicle Charging Stations produced by Sebastian Stein about our research. We also published our dataset.
In 2017 (April-May) I was a visiting scholar at Duke University with Vincent Conitzer.
As is common in AI and computer science, I publish my main fundamental contributions in the top (A*) conferences in my field: AAAI, AAMAS, IJCAI, ICAPS (see e.g. the core ranking or Google Scholar metrics). Contributions to operations research or other fields of science or engineering I publish in the respective journals. Below I've first highlighted some of my more recent publications on each of the three main challenges I'm working on. Below then can all my publications from the TU Delft research portal be found. Most are also included in the DBLP here. The website with publications of Mathijs de Weerdt at Google Scholar additionally contains an up-to-date citation count and h-index.
- Robustness and uncertainty:
Greg Neustroev, Mathijs de Weerdt (2020), Generalized Optimistic Q-Learning with Provable Efficiency, Bo An, Neil Yorke-Smith, Amal El Fallah Seghrouchni, Gita Sukthankar (Eds.), In Proceedings of AAMAS'20 p.913-921, International Foundation for Autonomous Agents and Multiagent Systems (IFAAMAS).
Greg Neustroev, Mathijs de Weerdt, Remco Verzijlbergh (2019), Discovery of Optimal Solution Horizons in Non-Stationary Markov Decision Processes with Unbounded Rewards, J. Benton, N. Lipovetzky, E. Onaindia, D.E. Smith, S. Srivastava (Eds.), In Proceedings of the Twenty-Ninth International Conference on Automated Planning and Scheduling Volume 29 p.292-300, Association for the Advancement of Artificial Intelligence (AAAI).
Frits de Nijs, Matthijs T.J. Spaan, Mathijs M. de Weerdt (2018), Preallocation and Planning under Stochastic Resource Constraints, S. McIlraith , K. Weinberger (Eds.), In Proceedings of the 32th AAAI Conference on Artificial Intelligence p.4662-4669, Association for the Advancement of Artificial Intelligence (AAAI). See also my blog post on Preallocation and Planning under Stochastic Resource Constraints.
Mathijs de Weerdt, Robert Baart, Lei He (2021), Single-machine scheduling with release times, deadlines, setup times, and rejection, In European Journal of Operational Research Volume 291 p.629-639. See also my blog post on What have satellite scheduling, the selected traveling salesperson (orienteering) problem, and make-to-order manufacturing in common?.
Pim van den Bogaerdt, Mathijs de Weerdt (2019), Lower Bounds for Uniform Machine Scheduling Using Decision Diagrams, Louis-Martin Rousseau, Kostas Stergiou (Eds.), In Integration of Constraint Programming, Artificial Intelligence, and Operations Research p.565-580, Springer.
Leon Planken, Mathijs de Weerdt, Roman van der Krogt (2012), Computing all-pairs shortest paths by leveraging low treewidth, In The Journal of Artificial Intelligence Research Volume 43 p.353-388.
- Multiple parties:
Mathijs de Weerdt, Michael Albert, Vincent Conitzer, Koos van der Linden (2018), Complexity of Scheduling Charging in the Smart Grid , In Proceedings of the Twenty-Seventh International Joint Conference on Artificial Intelligence (IJCAI-18) p.4736-4742, International Joint Conferences on Artifical Intelligence (IJCAI).
Frits De Nijs, Erwin Walraven, Mathijs M. De Weerdt, Matthijs T.J. Spaan (2017), Bounding the probability of resource constraint violations in multi-agent MDPs , In Proceedings of the 31st Conference on Artificial Intelligence, AAAI 2017 p.3562-3568, American Association for Artificial Intelligence (AAAI).
Mathijs de Weerdt, Paul Harrenstein, Vincent Conitzer (2014), Strategy-proof contract auctions and the role of ties, In Games and Economic Behavior Volume 86 p.405-420.
Other fields of science or engineering
- Road maintenance planning:
Joris Scharpff, Daan Schraven, Leentje Volker, Matthijs T.J. Spaan, Mathijs M. de Weerdt (2020), Can multiple contractors self-regulate their joint service delivery?: A serious gaming experiment on road maintenance planning, In Construction Management and Economics p.1-18.
- Transportation: routing and charging electric cars
M.M. de Weerdt, Sebastian Stein, Enrico Gerding, Valentin Robu, Nick Jennings (2015), Intention-Aware Routing of Electric Vehicles, In IEEE Transactions on Intelligent Transportation Systems Volume 17 p.1472 - 1482. See also my blog post on Solving Road Congestion Problems with Algorithms and this fun video made by Sebastian.
- Algorithms for railway operations
Zomer, J., Bešinović, N., de Weerdt, M. M., & Goverde, R. M. (2021). The Maintenance Scheduling and Location Choice Problem for Railway Rolling Stock. arXiv preprint arXiv:2103.00454.
Mulderij, J., Huisman, B., Tönissen, D., van der Linden, K., & de Weerdt, M. (2020). Train Unit Shunting and Servicing: a Real-Life Application of Multi-Agent Path Finding. arXiv preprint arXiv:2006.10422.
Jesse Mulderij, Jacobus van der Linden, Bob Huisman, Joris den Ouden, Marjan van den Akker, Han Hoogeveen and Mathijs de Weerdt (2021). TORS: a Train Unit Shunting and Servicing Simulator. AAMAS 2021.
- Electricity grid management:
Frits de Nijs, Mathijs M. de Weerdt, Matthijs T. J. Spaan (2019), Multi-agent Planning Under Uncertainty for Capacity Management, Peter Palensky, Miloš Cvetković, Tamás Keviczky (Eds.), In Intelligent Integrated Energy Systems p.197-213, Springer.
- Electricity markets:
Rens Philipsen, Germán Morales-España, Mathijs de Weerdt, Laurens De Vries (2019), Trading power instead of energy in day-ahead electricity markets, In Applied Energy Volume 233-234 p.802-815. See also my blog post summarizing this paper.
Germán Morales-España, Álvaro Lorca, Mathijs M. de Weerdt (2018), Robust unit commitment with dispatchable wind power, In Electric Power Systems Research Volume 155 p.58-66. See also my blog post on Stochastic Robust Optimization for Unit Commitment with Wind Curtailment.
Koos van der Linden, Mathijs de Weerdt, German Morales-Espana (2018), Optimal non-zero Price Bids for EVs in Energy and Reserves Markets using Stochastic Optimization, In 15th International Conference on the European Energy Market (EEM) p.1-5, IEEE. See my blog post on LinkedIn. This method is included in a general benchmarking tool for flexible-load electricity trading called B-FELSA, which is available on github.
- Order acceptance and scheduling:
Lei He, Arthur Guijt, Mathijs de Weerdt, Lining Xing, Neil Yorke-Smith (2019), Order Acceptance and Scheduling with Sequence-Dependent Setup Times: A new memetic algorithm and benchmark of the state of the art , In Computers and Industrial Engineering Volume 138 p.1-15.
Lei He, Mathijs de Weerdt, Neil Yorke-Smith (2019), Tabu-Based Large Neighbourhood Search for Time/Sequence-Dependent Scheduling Problems with Time Windows , J. Benton, N. Lipovetzky, E. Onaindia, D.E. Smith, S. Srivastave (Eds.), In Proceedings of the 29th International Conference on Automated Planning and Scheduling (ICAPS'19) Volume 29 p.186-194, Association for the Advancement of Artificial Intelligence (AAAI).
Lei He, Mathijs de Weerdt, Neil Yorke-Smith (2019), Time/Sequence-Dependent Scheduling: The design and evaluation of a general purpose tabu-based adaptive large neighbourhood search algorithm , In Journal of Intelligent Manufacturing p.1-28.
(ordered by year)
I'm responsible for the following bachelor courses in the core of the bachelor programme on Computer Science and Engineering in Delft.
- Algorithm Design (responsible teacher): this second-year course is on algorithmic techniques such as greedy algorithms, divide & conquer, dynamic programming and network flow; see also my open courseware version on Algorithm Design (recordings in Dutch, slides in English). I have designed this course from scratch, replacing a course on discrete mathematics. In more recent year I've tried to make this course more accessible to a broader range of students by facilitating different study paths through the introduction of what we call skill circuits. In these, students can select their own series of tasks to train themselves. This course is also a pioneering user of the virtual programming environment developed by Eelco Visser's group, called WebLab. Using this environment, students can receive automated feedback on the correctness of the implementation of their algorithms.
- Research Project (coordinator/responsible teacher): Based on the outcome of my 2018 curriculum-update committee work, I proposed a new course design for the final (third year, 15 EC) project of the CSE bachelor to let students learn some research skills. I have run pilots with 10, 15 and 45 students to prepare for the official 2020/2021 run of the course with close to 300 students.
- As a supervisor of one or two groups of students I'm regularly involved in the Bachelor Seminar, the Software Project (formerly known as BEP), the Honours Program (specifically on competitive programming), and of course the Research Project as well.
In the master I'm responsible for the courses offered by the Algorithmics group. I've delegated some of this responsibility to my colleagues, but I am still very much involved in the following master's courses.
- Advanced Algorithms (5 EC) (responsible teacher): I coordinate this course and am responsible for teaching one third of it, about exact algorithms for NP-hard problems.
- Algorithms for Intelligent Decision Making (5 EC) (co-teacher): I teach a part on algorithmic game theory.
- Algorithms for Intelligent Decision Making Project (5 EC) (co-teacher): here I supervise some of the groups and chair some of the sessions. In the groups students execute a small research project resulting in a paper.
- Master thesis project supervision (45 EC): I have supervised over 30 students on topics such as planning algorithms, and auction/mechanism design for multi-agent systems. I have also acted many times as an external committee member in other graduation committees. One of my students (Stranders) received the prize for the best Master’s thesis in Computer Science in 2006 from the Koninklijke Hollandsche Maatschappij der Wetenschappen for his thesis on Argumentation Based Decision Making for Trust in Multi- Agent Systems. Others worked for example in collaboration with Grontmij (now Sweco), Calendar42, ORTEC, SystemsNavigator, NS, and WithTheGrid. On average every year one of my students publishes a paper in a top venue with me based on his/her master thesis work.
- From 2016/2017 to 2020 I have been teaching Algorithm Design for teachers in secondary education as part of Inf4All. Please see my lecture on Greedy and Dynamic Programming algorithms on YouTube (in Dutch), and the related slides of this third meeting
- I was very active in the design of the 2018 CSE Bachelor curriculum with the following main changes: more freedom for students to personalize their program (electives in year 3), more research-oriented (e.g. the final project is now a research project), a focus on responsible computer science (throughout a number of courses), and a bit more attention to artificial intelligence.
- I was elected teacher of the year of Computer Science Delft in 2015.
- I was president of board of studies of Computer Science (OCI) from 2010-2014. The president sets the agenda, leads the discussions, and regularly communicates with the Director of Studies.
- Since 2019 I'm involved in the introduction and development of more courses on Artificial Intelligence in Delft, such as the AI Technology track in the master, and an AI "mini-minor" of 15 EC for master students with a non-CS master.
- I've given master classes on Smart Algorithms for Smart Grids for KIVI and Nyenrode/Allliander (2017).
- I've given tutorials on multi-agent planning and mechanism design at major conferences and the European Agent Systems Summer School.