Michel Reniers and Anton Wijs received funding from the Open Competition Domain Science-M program
Michel Reniers and Anton Wijs receive NWO funding to accelerate control software synthesis using GPUs for safer, faster system design.
Many devices and systems are controlled by software. Think of elevators, printers, cars, planes, bridges, and assembly lines in factories. Correct control is often crucial for safety. Supervisory Controller Synthesis is a technique to automatically derive control software, making the software correct by construction. The involved calculations, however, currently limit its applicability. In this project, Michel Reniers (Associate Professor Mechanical Engineering) and Anton Wijs (Associate Professor Mathematics and Computer Science) conduct research on how graphics processors, with an enormous computational power, can be used to scale up thistechnique, making it possible to generate more complex software in a shorter time span. Their proposal received funding from the Open Competition Domain Science-M program for two PhD students and one scientific programmer.
GUESS: GPU Enhanced Synthesis of Supervisory controllers
Modern high-tech systems are complex, and ensuring they behave safely requires advanced control software. Supervisory Control Synthesis (SCS) is a method that automatically generates such software based on system models and safety requirements. However, when applied to large, real-world systems, SCS runs into serious computational limits due to the massive number of possible system states. The GUESS project addresses this challenge by using the power of graphics processors (GPUs) to speed up and scale symbolic computations, particularly through Binary Decision Diagrams (BDDs).
By developing new algorithms for GPU-accelerated BDD manipulation, GUESS is pushing the boundaries of what鈥檚 possible in SCS, making it faster and more scalable. These innovations will also benefit other fields that rely on symbolic reasoning, such as formal verification, AI planning, circuit design, and reliability engineering. The techniques will be openly shared and are expected to impact both academic research and industry, with active collaborations already in place with experts across Europe. The project even looks ahead to cloud-based solutions, enabling large-scale computations using platforms like Amazon Web Services.
GUESS project beyond the lab
Beyond the lab, GUESS contributes to the safe and efficient control of infrastructure like bridges, tunnels, and water systems, making their development more reliable and affordable. The project supports a growing engineering approach called Synthesis-Based Engineering, which combines models and automation to create correct-by-design control software. Through its links with industry leaders such as ASML and national partners in the ESCET project, GUESS is helping to bring cutting-edge scientific advances directly into real-world applications.
Please read more at the NWO .