Complex and challenging decisions: the STAR Lab team explores trustworthy fundamental AI to help people make better decisions in situations from sustainable logistics, to robust planning under uncertainty, to fair energy markets.

Research

The TU Delft STAR Lab confronts a fundamental question of the AI era: how can machine intelligence assist individuals and groups who face complex decision trade-offs? We research how bringing together data and models, human preferences, and AI reasoning can facilitate outcomes better for society. We make impact through partnering with companies, universities, municipalities and government departments.
Data-driven optimisation

We bring together data and models to solve complex combinatorial optimisation problems. By tackling real-world problems in socio-technical situations, we gain fundamental insights into the complementary interplay of machine learning and model-based search.
Agent-based social simulation

We design scalable computational models using micro-simulation, behavioural and spatial data, and participatory modelling, to understand the trade-offs of policy interventions, particularly in urban settings.
Learning & scheduling under uncertainty

We recognise that decision making is ripe with uncertainty, and this uncertainty has multiple forms. By disentangling aleatoric and epistemic uncertainty, we develop understanding of how AI algorithms can be made more reliable.

Projects

FMaaS
FMaaS
NWO (2024-29)

Freight Management as a Service

Optimisation

Uncertainty


Learn more
AIFES
AIFES
AiNed (2023)

AI for the future energy system

Optimisation

Uncertainty


Learn more
Plug-In
Plug-In
Dutch Ministry EZK (2022-26)

HGV energy hubs

Optimisation

Uncertainty


Learn more
TULIPS logo
TULIPS
EU Green Deal (2022-26)

Sustainable landside transport

Optimisation

Simulation


Learn more
E-pi logo
Epistemic AI
Horizon 2020 (2021-26)

Redefining the basis for AI

Uncertainty


Learn more
B4B logo
Brains for Buildings
Dutch Ministry EZK (2021-25)

Data-driven building optimisation

Uncertainty

Optimisation


Learn more
Voestalpine logo
Voestalpine
Industrial (2021-22)

Job planning and sequencing

Optimisation


Learn more
OPTIMAL
OPTIMAL
NWO (2020-25)

Machine learning for optimisation

Optimisation


Learn more
TAILOR logo
TAILOR
EU Horizon 2020 (2020-24)

Trustworthy AI

Simulation

Optimisation


Learn more
SAM-FMS
SAM-FMS
NWO (2020-24)

Scheduling cyber-physical systems

Optimisation


Learn more
HaPSISH
HaPSISH
NWO (2016-20)

Energy system integration

Optimisation

Uncertainty


Learn more
DCSMART
DCSMART
EU Horizon 2020 (2016-19)

Energy system integration

Simulation

Uncertainty


Learn more

Recent News

Accepted paper

Journal of Process Control, "Data-Driven Robust Optimization with Machine Learning Enabled Uncertainty Set"

Accepted paper

BNAIC 2024, "An Efficient Decremental Algorithm for Simple Temporal Networks"

Master's graduate

Laurens Krudde, MSc, graduated on "Demand Responsive Transport to Replace a Fixed-Line Bus Service"

New STAR Lab doctoral student

Xueying Ou, FMaaS project

Accepted paper

Mathematical Programming, "Machine Learning Augmented Branch and Bound for Mixed Integer Linear Programming" (pdf)

Master's graduate

Michel Woo, MSc, graduated on "A Hybrid Genetic Search Approach to Optimizing Last-Mile Vehicle Routing"

Master's graduate

Siddhartha Sen, MSc, graduated on "Golf Course Routing Using Artificial Intelligence"

Master's graduate

Vincent Bockstael, MSc, graduated cum laude on "Spectral Modularity: Conceptual Challenges, Algorithmic Enhancements and Systematic Benchmarking"

Master's graduate

Martin Michaux, MSc, graduated on "A Scalable Knowledge Graph-Powered System for Multi-Document Query-Answering"

Master's graduate

Lars van Koetsveld van Ankeren, MSc, graduated cum laude on "Real-time Monitoring of Building Energy Systems"

Bachelor's graduate

Bertold Kovacs, BSc, graduated from the honours programme

Bachelor's graduate

Yehor Kozyr, BSc, graduated from the honours programme

Accepted paper

ITEC 2024, "Multiobjective System Sizing for Heavy-Duty Electric Vehicle Charging Stations"

Master's graduate

Jeroen Groenheide, MSc, graduated on "On Learning for Node Selection in the Branch-and-Bound Algorithm using Reinforcement Learning"

Master's graduate

Octav Pocola, MSc, graduated on "Optimal Battery Sizing and Control in a Stacked Revenue Model Incorporating a Renewable Energy Community"

Accepted paper

EWRL 2024, "Epistemic Bellman Operators"

Accepted paper

SMS 2024, "Enhancing Predictions for Self-Compacting Mortar Properties with Multilayer Perceptron and Data Augmentation"

Master's graduate

Tim Huisman, MSc, graduated on "Algorithmic Solutions for Improved Carrier-Shipper Matching in a Competitive Transport Marketplace"

Master's graduate

Luca Cras, MSc, graduated on "Evaluating the Impact and Opportunities of Physics-Informed Machine Learning on the Task of Greenhouse Humidity Prediction"

Accepted paper

INFORMS Journal on Computing, "Multi-Objective Linear Ensembles for Robust and Sparse Training of Few-Bit Neural Networks" (pdf)

Master's graduate

Isa Rethans, MSc, graduated on "Predictive Analysis and Key Drivers for PostNL's Cost Per Package"

New STAR Lab doctoral student

Joost Commandeur, Shell e-Mobility

Accepted paper

TRC-30 symposium, "Psychological Factors in Travel Behaviour Interpretation with Social Media Data"

Accepted paper

DSO 2024 workshop at IJCAI'24, "Neural Decision Diagrams for Job Shop Scheduling"

Accepted paper

ISMP 2024, "Machine Learning Augmented Branch and Bound for Mixed Integer Linear Programming" (journal pdf)

Accepted paper

ISMP 2024, "Slimming Down and Beefing Up: Multi-Objective Optimization and Training for Few-Bit Neural Networks"