University of Warwick

There is a growing interest in the study of periodic phenomena in

large-scale nonlinear dynamical systems. Often the high-dimensional

system has only low-dimensional dynamics, e.g., many reaction-diffusion

systems or Navier-Stokes flows at low Reynolds number. We present an

approach that exploits this property in order to compute branches of

periodic solutions of the large system of ordinary differential

equations (ODEs) obtained after a space discretisation of the PDE. We

call our approach the Newton-Picard method. Our method is based on the

recursive projection method (RPM) of Shroff and Keller but extends this

method in many different ways. Our technique tries to combine the

performance of straightforward time integration with the advantages of

solving a nonlinear boundary value problem using Newton's method and a

direct solver. Time integration works well for very stable limit

cycles. Solving a boundary value problem is expensive, but works also

for unstable limit cycles.

\\

\\

We will present some background material on RPM. Next we will explain

the basic features of the Newton-Picard method for single shooting. The

linearised system is solved by a combination of direct and iterative

techniques. First, we isolate the low-dimensional subspace of unstable

and weakly stable modes (using orthogonal subspace iteration) and

project the linearised system on this subspace and on its

(high-dimensional) orthogonal complement. In the high-dimensional

subspace we use iterative techniques such as Picard iteration or GMRES.

In the low-dimensional (but "hard") subspace, direct methods such as

Gaussian elimination or a least-squares are used. While computing the

projectors, we also obtain good estimates for the dominant,

stability-determining Floquet multipliers. We will present a framework

that allows us to monitor and steer the convergence behaviour of the

method.

\\

\\

RPM and the Newton-Picard technique have been developed for PDEs that

reduce to large systems of ODEs after space discretisation. In fact,

both methods can be applied to any large system of ODEs. We will

indicate how these methods can be applied to the discretisation of the

Navier-Stokes equations for incompressible flow (which reduce to an

index-2 system of differential-algebraic equations after space

discretisation when written in terms of velocity and pressure.)

\\

\\

The Newton-Picard method has already been extended to the computation

of bifurcation points on paths of periodic solutions and to multiple

shooting. Extension to certain collocation and finite difference

techniques is also possible.

In many applications of interest, such as the conformational

dynamics of molecules, large deterministic systems can exhibit

stochastic behaviour in a relative small number of coarse-grained

variables. This kind of dimension reduction, from a large deterministic

system to a smaller stochastic one, can be very useful in understanding

the problem. Whilst the subject of statistical mechanics provides

a wealth of explicit examples where stochastic models for coarse

variables can be found analytically, it is frequently the case

that applications of interest are not amenable to analytic

dimension reduction. It is hence of interest to pursue computational

algorithms for such dimension reduction. This talk will be devoted

to describing recent work on parameter estimation aimed at

problems arising in this context.

\\

\\

Joint work with Raz Kupferman (Jerusalem) and Petter Wiberg (Warwick)

Plane Couette flow - the flow between two infinite parallel plates moving in opposite directions -

undergoes a discontinuous transition from laminar flow to turbulence as the Reynolds number is

increased. Due to its simplicity, this flow has long served as one of the canonical examples for understanding shear turbulence and the subcritical transition process typical of channel and pipe flows. Only recently was it discovered in very large aspect ratio experiments that this flow also exhibits remarkable pattern formation near transition. Steady, spatially periodic patterns of distinct regions of turbulent and laminar flow emerges spontaneously from uniform turbulence as the Reynolds number is decreased. The length scale of these patterns is more than an order of magnitude larger than the plate separation. It now appears that turbulent-laminar patterns are inevitable intermediate states on the route from turbulent to laminar flow in many shear flows. I will explain how we have overcome the difficulty of simulating these large scale patterns and show results from studies of three types of patterns: periodic, localized, and intermittent.

Isostatic mounts are used in applications like telescopes and robotics to move and hold part of a structure in a desired pose relative to the rest, by driving some controls rather than driving the subsystem directly. To achieve this successfully requires an understanding of the coupled space of configurations and controls, and of the singularities of the mapping from the coupled space to the space of controls. It is crucial to avoid such singularities because generically they lead to large constraint forces and internal stresses which can cause distortion. In this paper we outline design principles for isostatic mount systems for dynamic structures, with particular emphasis on robots.

In many applications it is of interest to compute minimizers of

a functional I(u) which is the of the form $J(u)=\Phi(u)+R(u)$,

with $R(u)$ quadratic. We describe a stochastic algorithm for

this problem which avoids explicit computation of gradients of $\Phi$;

it requires only the ability to sample from a Gaussian measure

with Cameron-Martin norm squared equal to $R(u)$, and the ability

to evaluate $\Phi$. We show that, in an appropriate parameter limit,

a piecewise linear interpolant of the algorithm converges weakly to a noisy

gradient flow. \\

Joint work with Natesh Pillai (Harvard) and Alex Thiery (Warwick).

I will describe recent work with Charles Fefferman on a

construction of families of analytic almost-sharp fronts for SQG. These

are special solutions of SQG which have a very sharp transition in a

very thin layer. One of the main difficulties of the construction is the

fact that there is no formal limit for the family of equations. I will

show how to overcome this difficulty, linking the result to joint work

with C. Fefferman and Kevin Luli on the existence of a "spine" for

almost-sharp fronts. This is a curve, defined for every time slice by a

measure-theoretic construction, that describes the evolution of the

almost-sharp front.

The symmetries of a dynamical system have a big effect on its typical behaviour. The most obvious effect is pattern formation - the dynamics itself may be symmetric, though often the symmetry of the system is 'broken', and the state has less symmetry than the system. The resulting phenomena are fairly well understood for steady and time-periodic states, and quite a bit can be said for chaotic dynamics. More recently, the concept of 'symmetry' has been generalised to address applications to physical and biological networks. One consequence is a new approach to patterns of synchrony and phase relations. The lecture will describe some of the high points of the emerging theories, including applications to evolution, locomotion, human balance and fluid dynamics.

Human behaviour can show surprising properties when looked at from a collective point of view. Data on collective behaviour can be gleaned from a number of sources, and mobile phone data are increasingly becoming used. A major challenge is combining behavioural data with health data. In this talk I will describe our approach to understanding behaviour change related to change in health status at a collective level.

I will show that one can (at least in theory) guarantee the "validity" of a numerical approximation of a solution of the 3D Navier-Stokes equations using an explicit a posteriori test, despite the fact that the existence of a unique solution is not known for arbitrary initial data.

The argument relies on the fact that if a regular solution exists for some given initial condition, a regular solution also exists for nearby initial data ("robustness of regularity"); I will outline the proof of robustness of regularity for initial data in $H^{1/2}$.

I will also show how this can be used to prove that one can verify numerically (at least in theory) the following statement, for any fixed R > 0: every initial condition $u_0\in H^1$ with $\|u\|_{H^1}\le R$ gives rise to a solution of the unforced equation that remains regular for all $t\ge 0$.

This is based on joint work with Sergei Chernysehnko (Imperial), Peter Constantin (Chicago), Masoumeh Dashti (Warwick), Pedro Marín-Rubio (Seville), Witold Sadowski (Warsaw/Warwick), and Edriss Titi (UC Irivine/Weizmann).

Tonelli gave the first rigorous treatment of one-dimensional variational problems, providing conditions for existence and regularity of minimizers over the space of absolutely continuous functions.  He also proved a partial regularity theorem, asserting that a minimizer is everywhere differentiable, possible with infinite derivative, and that this derivative is continuous as a map into the extended real line.  Some recent work has lowered the smoothness assumptions on the Lagrangian for this result to various Lispschitz and H\"older conditions.  In this talk we will discuss the partial regularity result, construct examples showing that mere continuity of the Lagrangian is an insufficient condition.