Learning to Adapt
Robbert Reijnen鈥檚 research shows how deep reinforcement learning can dynamically optimize machine scheduling, helping industries respond to evolving challenges and complex decision-making environments.
Changing Landscapes
Industries increasingly face decision-making problems that involve balancing multiple objectives under tight constraints. In manufacturing, logistics, and energy systems, scheduling tasks efficiently is critical. Reijnen鈥檚 research focuses on combinatorial optimization problems, which are known for their vast and structured solution spaces. Traditional algorithms often rely on fixed configurations, making them vulnerable when the problem landscape shifts.
Learning Algorithms
Reijnen proposes a learning-based approach using deep reinforcement learning to dynamically adjust algorithm parameters and operator selection during the optimization process. This method treats algorithm control as a sequential decision-making problem, allowing the system to respond to real-time feedback and adapt its strategy. His work shows that dynamic control can help algorithms escape local optima and explore more promising regions of the solution space.
Practical Impact
The research offers insights for sectors that rely on efficient scheduling, such as manufacturing, transport, and energy. Businesses often struggle with static systems that fail to adapt to disruptions or changing demands. Reijnen鈥檚 frameworks demonstrate how adaptive algorithms can support more resilient and responsive operations.
果冻传媒 Expertise
Within the Information Systems group, researchers combine optimization, data science, and machine learning to tackle real-world challenges. Reijnen鈥檚 work contributes to this effort by bridging algorithm design with practical applications in dynamic environments.
Robbert Reijnen defended his thesis on October 13, 2025. Title of the thesis: Supervisors: Yingqian Zhang and Zaharah Bukhsh.