Some of our research projects
Robust Optimization
Robust Optimization (RO) is a powerful modeling methodology that addresses optimization problems in the presence of uncertain data. Unlike traditional optimization, which assumes exact parameter values, RO explicitly considers the uncertainty associated with the data. This makes it particularly relevant for real-world scenarios where data can be imprecise or subject to variation. RO aims to find solutions that remain feasible even when the data deviates from its nominal values. It ensures stability by considering worst-case scenarios, preventing solutions from becoming infeasible due to small perturbations in the data. In the past decade, RO has shown a great importance in solving practical problems, including but not limited to, supply chain, network design, disruption management, transportation, circularity, and data-science. In our group, we both work on the theory of (distributionally) robust optimization and use robust optimization methodology for solving various applied research problems.
Optimization Under Uncertainty
Optimization under uncertainty is an umbrella term for various optimization tools (including mathematical programming) used for stochastic sequential decision problems. From our mathematical programming perspective, we are interested in how to best-combine classic Operations Research approaches with learning-based mechanisms. We utilize tools from stochastic programming, Monte-Carlo simulation, and reinforcement learning, seeking to integrate them for various real-world use cases.
Transportation and Logistics
Transportation and logistics problems are fundamentally combinatorial optimization problems that require the art of mathematical programming navigate their vast decision spaces. Within this context, we have extensive experience in addressing deterministic problems using both exact and heuristic approaches such as column generation, branch-and-price-and-cut, Bender鈥檚 decomposition, metaheuristics, or a combination of these methods. Our philosophy is to start from practice, acquire domain knowledge about transport and logistics use-cases, and identify fundamental challenges that require scientific research. For example, the AI4ROAD project focusses on developing real-time planning tools for our industry partner Van der Wal Transport. Researchers and PhD students collaborate to develop innovative exact and heuristic appmathematical porroaches for road-transport planning in the Netherlands.
Involved researchers
Ahmadreza Marandi, Albert Schrotenboer, Tom van Woensel, Rolf van Lieshout Layla Martin, Nico Dellaert and Mehrdad Mohammadi
More Information
For more information, please feel free to reach out to Albert Schrotenboer