David Stewart is a known expert in computational methods, focusing on optimization and differential equations. His research delves into the complexities of nonsmooth and discontinuous dynamics, addressing challenges in mechanical impacts (such as bouncing balls), and electronic systems with switching elements (like diodes). In the realm of optimization, Professor Stewart explores numerical methods and optimal control, aiming to develop efficient algorithms for complex mathematical problems. 

More recently, Professor Stewart has expanded his research interests to include machine learning. He investigates how well functions can be approximated by neural networks and the optimization of weights in autoencoders, bridging the gap between traditional computational methods and modern AI techniques.

Research interests

• Numerical analysis
• Mathematical modeling
• Optimization
• Differential equations