Physics-Informed Deep Operator Network (PI-DeepONet) for Stabilization of Coupled PDE-ODE Systems

Coupled PDE-ODE systems are commonly used to model complex physical processes, but designing controllers is challenging due to nonlinear dynamics. Here, we propose a Physics-Informed Deep Operator Network (PI-DeepONet) that directly learns backstepping kernel functions from data, ensuring consistency with the system's physical laws. These kernels are used to design controllers that ensure exponential stabilization without requiring full-state measurements or analytical solutions. Simulation results show that the approach is accurate, reliable, and practically applicable.

کلیدواژه ها:

PDE-ODE systems ، backstepping control ، neural operator ، PI-Deep ONet ، stabilization ، data-driven control