Bounding extrema over global attractors using polynomial optimisation

Abstract

We describe a framework for bounding extreme values of quantities on global attractors of differential dynamical systems. A global attractor is the minimal set that attracts all bounded sets; it contains all forward-time limit points. Our approach uses (generalised) Lyapunov functions to find attracting sets, which must contain the global attractor, and the choice of Lyapunov function is optimised based on the quantity whose extreme value one aims to bound. We also present a non-global framework for bounding extrema over the minimal set that is attracting in a specified region of state space. If the dynamics are governed by ordinary differential equations, and the equations and quantities of interest are polynomial, then our methods can be implemented computationally using polynomial optimisation. In particular, we enforce nonnegativity of certain polynomial expressions by requiring them to be representable as sums of squares, leading to a convex optimisation problem that can be recast as a semidefinite program and solved computationally. This computer assistance lets one construct complicated polynomial Lyapunov functions. Computations are illustrated using three examples. The first is the chaotic Lorenz system, where we bound extreme values of various monomials of the coordinates over the global attractor. In the second example we bound extreme values in a nine-mode truncation of fluid dynamics which displays long-lived chaotic transients. The third example has two locally stable limit cycles, each with its own basin of attraction, and we apply our non-global framework to construct bounds for one basin that do not apply to the other. For each example we compute Lyapunov functions of polynomial degrees up to at least eight. In cases where we can judge the sharpness of our bounds, they are sharp to at least three digits when the polynomial degree is at least four or six.

arXiv comment25 pages; v4: minor additions and changes to text

References in corpus (9)

• Bounds for deterministic and stochastic dynamical systems using sum-of-squares optimization
• Global Stability Analysis of Fluid Flows using Sum-of-Squares
• Optimal bounds and extremal trajectories for time averages in nonlinear dynamical systems
• Integrals of motion and the shape of the attractor for the Lorenz model
• Bounds on mean energy in the Kuramoto-Sivashinsky equation computed using semidefinite programming
• Bounding averages rigorously using semidefinite programming: mean moments of the Lorenz system
• Automatic Generation of Bounds for Polynomial Systems with Application to the Lorenz System
• Finding extremal periodic orbits with polynomial optimisation, with application to a nine-mode model of shear flow
• Sparse sum-of-squares (SOS) optimization: A bridge between DSOS/SDSOS and SOS optimization for sparse polynomials

Cited by in corpus (6)

• Chordal and factor-width decompositions for scalable semidefinite and polynomial optimization
• Auxiliary Functions as Koopman Observables: Data-Driven Analysis of Dynamical Systems via Polynomial Optimization
• A study of the double pendulum using polynomial optimization
• Minimum wave speeds in monostable reaction-diffusion equations: sharp bounds by polynomial optimization
• Uncertainty propagation for nonlinear dynamics: A polynomial optimization approach
• Data-driven nonlinear aerodynamics models with certifiably optimal boundedness properties