Abstract:

Understanding physical phenomena and developing mathematical models to approximate them is at the core of natural sciences. These models must be interpretable and provide reasonable predictability to be useful. This is usually done either by physical reasoning or data-driven modeling. In this talk, a systematic data-driven identification method using LDDE operators is presented and applied to both LDDE and known nonlinear systems. The delays – in the developed method – are identified by the method itself and not left to the user. They are assumed to be inherent in the dynamics and not merely new coordinates and sampling techniques for approximation. The talk is more toward the identification method and approximating space construction. LDDE has an infinite-dimension dynamical space. Hopefully, this linear depiction would provide more physical insights and local reasoning. In real-life, interactions have finite traveling speeds, and hence dynamical systems should experience retardations.

Biography:

. 

________________________________________

All faculty, researchers and students are invited to attend.