AIMS Lab Seminar - Adam Sliwiak - Sensitivity analysis of chaotic dynamical systems: space-split sensitivity (S3) method
Title: Sensitivity analysis of chaotic dynamical systems: space-split sensitivity (S3) method
Abstract: Derivatives of statistical quantities describing chaos are highly desirable in science and engineering. They play a major role in climate prediction and control, turbulent flow analysis, and design optimization. Due to the well-known butterfly effect, however, sensitivity analysis of chaotic systems is a particularly challenging task from the computational point of view.
In this talk, I present a novel method, called the space-split sensitivity (S3), derived based on the ergodic dynamical systems theory. S3 is a computable reformulation of Ruelle’s linear response formula, which is a rigorous closed-form expression for the derivative of long-time averages in uniformly hyperbolic systems. The S3 formula can be easily translated to efficient trajectory-based algorithms that provably converge as a typical Monte Carlo method.
Joint work with Prof. Qiqi Wang (MIT AeroAstro) and Nisha Chandramoorthy (MIT Mechanical Engineering).
Date/Time: Monday March 29, 11:30am - 12:20pm
Join Zoom Meeting
Meeting ID: 910 5698 2661