
One way of getting insight into nonGaussian measures, posed on infinite
dimensional Hilbert spaces, is to first obtain best fit Gaussian
approximations, which are more amenable to numerical approximation. These
Gaussians can then be used to accelerate sampling algorithms. This begs the
questions of how one should measure optimality and how the optimizers can be
obtained. Here, we consider the problem of minimizing the distance with respect
to relative entropy. We examine this minimization problem by seeking roots of
the first variation of relative entropy, taken with respect to the mean of the
Gaussian, leaving the covariance fixed. Adapting a convergence analysis of
RobbinsMonro to the infinite dimensional setting, we can justify the
application of this algorithm and highlight necessary assumptions to ensure
convergence, not only in the context of relative entropy minimization, but
other infinite dimensional problems as well. Numerical examples in path space,
showing the robustness of this method with respect to dimension, are provided.

We examine a degenerate, dispersive, nonlinear wave equation related to the
evolution of partially molten rock in dimensions two and higher. This
simplified model, for a scalar field capturing the melt fraction by volume, has
been studied by direct numerical simulation where it has been observed to
develop stable solitary waves. In this work, we prove local in time
wellposedness results for the time dependent equation, on both the whole space
and the torus, for dimensions two and higher. We also prove the existence of
the solitary wave solutions in dimensions two and higher.

Metastable condensed matter typically fluctuates about local energy minima at
the femtosecond time scale before transitioning between local minima after
nanoseconds or microseconds. This vast scale separation limits the
applicability of classical molecular dynamics methods and has spurned the
development of a host of approximate algorithms. One recently proposed method
is diffusive molecular dynamics which aims to integrate a system of ordinary
differential equations describing the likelihood of occupancy by one of two
species, in the case of a binary alloy, while quasistatically evolving the
locations of the atoms. While diffusive molecular dynamics has shown to be
efficient and provide agreement with observations, it is fundamentally a model,
with unclear connections to classical molecular dynamics.
In this work, we formulate a spindiffusion stochastic process and show how
it can be connected to diffusive molecular dynamics. The spindiffusion model
couples a classical overdamped Langevin equation to a kinetic Monte Carlo model
for exchange amongst the species of a binary alloy. Under suitable assumptions
and approximations, spindiffusion can be shown to lead to diffusive molecular
molecular dynamics type models. The key assumptions and approximations include
a well defined time scale separation, a choice of spin exchange rates, a low
temperature approximation, and a mean field type approximation. We derive
several models from different assumptions and show their relationship to
diffusive molecular dynamics. Differences and similarities amongst the models
are explored in a simple test problem.

Weak turbulence is a phenomenon by which a system generically transfers
energy from low to high wave numbers, while persisting for all finite time. It
has been conjectured by Bourgain that the 2D defocusing nonlinear Schr\"odinger
equation (NLS) on the torus has this dynamic, and several analytical and
numerical studies have worked towards addressing this point.
In the process of studying the conjecture, Colliander, Keel, Staffilani,
Takaoka, and Tao introduced a "toy model" dynamical system as an approximation
of NLS, which has been subsequently studied numerically. In this work, we
formulate and examine several numerical schemes for integrating this model
equation. The model has two invariants, and our schemes aim to conserve at
least one of them. We prove convergence in some cases, and our numerical
studies show that the schemes compare favorably to others, such as Trapezoidal
Rule and fixed step fourth order RungeKutta. The preservation of the
invariants is particularly important in the study of weak turbulence as the
energy transfer tends to occur on long time scales.

We consider radial solutions to the Schr\"odingerPoisson system in three
dimensions with an external smooth potential with Coulomblike decay. Such a
system can be viewed as a model for the interaction of dark matter with a
bright matter background in the nonrelativistic limit. We find that there are
infinitely many critical points of the Hamiltonian, subject to fixed mass, and
that these bifurcate from solutions to the associated linear problem at zero
mass. As a result, each branch has a different topological character defined by
the number of zeros of the radial states. We construct numerical approximations
to these nonlinear states along the first several branches. The solution
branches can be continued, numerically, to large mass values, where they become
asymptotic, under a rescaling, to those of the Schr\"odingerPoisson problem
with no external potential. Our numerical computations indicate that the ground
state is orbitally stable, while the excited states are linearly unstable for
sufficiently large mass.

The Derivative Nonlinear Schr\"odinger equation is an $L^2$critical
nonlinear dispersive equation model for Alfv\'en waves in space plasmas. Recent
numerical studies on an $L^2$supercritical extension of this equation provide
evidence of finite time singularities. Near the singular point, the solution is
described by a universal profile that solves a nonlinear elliptic eigenvalue
problem depending only on the strength of the nonlinearity. In the present
work, we describe the deformation of the profile and its parameters near
criticality, combining asymptotic analysis and numerical simulations.

We prove that solitons (or solitary waves) of the ZakharovKuznetsov (ZK)
equation, a physically relevant high dimensional generalization of the
Kortewegde Vries (KdV) equation appearing in Plasma Physics, and having mixed
KdV and nonlinear Schr\"odinger (NLS) dynamics, are strongly asymptotically
stable in the energy space in the physical region. We also prove that the sum
of wellarranged solitons is stable in the same space. Orbital stability of ZK
solitons is wellknown since the work of de Bouard. Our proofs follow the ideas
by Martel and Martel and Merle, applied for generalized KdV equations in one
dimension. In particular, we extend to the high dimensional case several
monotonicity properties for suitable halfportions of mass and energy; we also
prove a new Liouville type property that characterizes ZK solitons, and a key
Virial identity for the linear and nonlinear part of the ZK dynamics, obtained
independently of the mixed KdVNLS dynamics. This last Virial identity relies
on a simple sign condition, which is numerically tested for the two and three
dimensional cases, with no additional spectral assumptions required. Possible
extensions to higher dimensions and different nonlinearities could be obtained
after a suitable local wellposedness theory in the energy space, and the
verification of a corresponding sign condition.

Diffusive molecular dynamics is a novel model for materials with atomistic
resolution that can reach diffusive time scales. The main ideas of diffusive
molecular dynamics are to first minimize an approximate variational Gaussian
free energy of the system with respect to the mean atomic coordinates
(averaging over many vibrational periods), and to then to perform a diffusive
step where atoms and vacancies (or two species in a binary alloy) flow on a
diffusive time scale via a master equation. We present a mathematical framework
for studying this algorithm based upon relative entropy, or KullbackLeibler
divergence. This adds flexibility in how the algorithm is implemented and
interpreted. We then compare our formulation, relying on relative entropy and
absolute continuity of measures, to existing formulations. The main difference
amongst the equations appears in a model for vacancy diffusion, where
additional entropic terms appear in our development.

We examine the Petviashvilli method for solving the equation $ \phi  \Delta
\phi = \phi^{p1} \phi$ on a bounded domain $\Omega \subset \mathbb{R}^d$
with Dirichlet boundary conditions. We prove a local convergence result, using
spectral analysis, akin to the result for the problem on $\mathbb{R}$ by
Pelinovsky & Stepanyants, 2004. We also prove a global convergence result by
generating a suite of nonlinear inequalities for the iteration sequence, and we
show that the sequence has a natural energy that decreases along the sequence.

Metastability is a common obstacle to performing long molecular dynamics
simulations. Many numerical methods have been proposed to overcome it. One
method is parallel replica dynamics, which relies on the rapid convergence of
the underlying stochastic process to a quasistationary distribution. Two
requirements for applying parallel replica dynamics are knowledge of the time
scale on which the process converges to the quasistationary distribution and a
mechanism for generating samples from this distribution. By combining a
FlemingViot particle system with convergence diagnostics to simultaneously
identify when the process converges while also generating samples, we can
address both points. This variation on the algorithm is illustrated with
various numerical examples, including those with entropic barriers and the 2D
LennardJones cluster of seven atoms.

In this paper we study algorithms to find a Gaussian approximation to a
target measure defined on a Hilbert space of functions; the target measure
itself is defined via its density with respect to a reference Gaussian measure.
We employ the KullbackLeibler divergence as a distance and find the best
Gaussian approximation by minimizing this distance. It then follows that the
approximate Gaussian must be equivalent to the Gaussian reference measure,
defining a natural function space setting for the underlying calculus of
variations problem. We introduce a computational algorithm which is
welladapted to the required minimization, seeking to find the mean as a
function, and parameterizing the covariance in two different ways: through low
rank perturbations of the reference covariance; and through Schr\"odinger
potential perturbations of the inverse reference covariance. Two applications
are shown: to a nonlinear inverse problem in elliptic PDEs, and to a
conditioned diffusion process. We also show how the Gaussian approximations we
obtain may be used to produce improved pCNMCMC methods which are not only
welladapted to the highdimensional setting, but also behave well with respect
to small observational noise (resp. small temperatures) in the inverse problem
(resp. conditioned diffusion).

The parallel replica dynamics, originally developed by A.F. Voter,
efficiently simulates very long trajectories of metastable Langevin dynamics.
We present an analogous algorithm for discrete time Markov processes. Such
Markov processes naturally arise, for example, from the time discretization of
a continuous time stochastic dynamics. Appealing to properties of
quasistationary distributions, we show that our algorithm reproduces exactly
(in some limiting regime) the law of the original trajectory, coarsened over
the metastable states.

In a variety of applications it is important to extract information from a
probability measure $\mu$ on an infinite dimensional space. Examples include
the Bayesian approach to inverse problems and possibly conditioned) continuous
time Markov processes. It may then be of interest to find a measure $\nu$, from
within a simple class of measures, which approximates $\mu$. This problem is
studied in the case where the KullbackLeibler divergence is employed to
measure the quality of the approximation. A calculus of variations viewpoint is
adopted and the particular case where $\nu$ is chosen from the set of Gaussian
measures is studied in detail. Basic existence and uniqueness theorems are
established, together with properties of minimising sequences. Furthermore,
parameterisation of the class of Gaussians through the mean and inverse
covariance is introduced, the need for regularisation is explained, and a
regularised minimisation is studied in detail. The calculus of variations
framework resulting from this work provides the appropriate underpinning for
computational algorithms.

We study a derivative nonlinear Schr\"{o}dinger equation, allowing
noninteger powers in the nonlinearity, $u^{2\sigma} u_x$. Making careful use
of the energy method, we are able to establish shorttime existence of
solutions with initial data in the energy space, $H^1$. For more regular
initial data, we establish not just existence of solutions, but also
wellposedness of the initial value problem. These results hold for realvalued
$\sigma\geq 1,$ while prior existence results in the literature require
integervalued $\sigma$ or $\sigma$ sufficiently large ($\sigma \geq 5/2$), or
use higherregularity function spaces.

We study soliton solutions to a generalized Korteweg  de Vries (KdV)
equation with a saturated nonlinearity, following the line of inquiry of the
authors for the nonlinear Schr\"odinger equation (NLS). KdV with such a
nonlinearity is known to possess a minimalmass soliton. We consider a small
perturbation of a minimalmass soliton and identify a system of ODEs, which
models the behavior of the perturbation for short times. This connects nicely
to a work of Comech, Cuccagna & Pelinovsky (2007). These ODEs form a simple
dynamical system with a single unstable hyperbolic fixed point with two
possible dynamical outcomes. A particular feature of the dynamics are that they
are nonoscillatory. This distinguishes the KdV problem from the analogous NLS
one.

We present a numerical study of a derivative nonlinear Schr\"odinger equation
with a general power nonlinearity, $\psi^{2\sigma}\psi_x$. In the
$L^2$supercritical regime, $\sigma>1$, our simulations indicate that there is
a finite time singularity. We obtain a precise description of the local
structure of the solution in terms of blowup rate and asymptotic profile, in a
form similar to that of the nonlinear Schr\"odinger equation with supercritical
power law nonlinearity.

We experimentally explore solutions to a model Hamiltonian dynamical system
derived in Colliander et al., 2012, to study frequency cascades in the cubic
defocusing nonlinear Schr\"odinger equation on the torus. Our results include a
statistical analysis of the evolution of data with localized amplitudes and
random phases, which supports the conjecture that energy cascades are a generic
phenomenon. We also identify stationary solutions, periodic solutions in an
associated problem and find experimental evidence of hyperbolic behavior. Many
of our results rely upon reframing the dynamical system using a hydrodynamic
formulation.

Parallel replica dynamics is a method for accelerating the computation of
processes characterized by a sequence of infrequent events. In this work, the
processes are governed by the overdamped Langevin equation. Such processes
spend much of their time about the minima of the underlying potential,
occasionally transitioning into different basins of attraction. The essential
idea of parallel replica dynamics is that the exit time distribution from a
given well for a single process can be approximated by the minimum of the exit
time distributions of $N$ independent identical processes, each run for only
1/Nth the amount of time.
While promising, this leads to a series of numerical analysis questions about
the accuracy of the exit distributions. Building upon the recent work in Le
Bris et al., we prove a unified error estimate on the exit distributions of the
algorithm against an unaccelerated process. Furthermore, we study a dephasing
mechanism, and prove that it will successfully complete.

We consider a derivative nonlinear Schr\"odinger equation with a general
nonlinearity. This equation has a two parameter family of solitary wave
solutions. We prove orbital stability/instability results that depend on the
strength of the nonlinearity and, in some instances, their velocity. We
illustrate these results with numerical simulations.

We consider an equation for a thinfilm of fluid on a rotating cylinder and
present several new analytical and numerical results on steady state solutions.
First, we provide an elementary proof that both weak and classical steady
states must be strictly positive so long as the speed of rotation is nonzero.
Next, we formulate an iterative spectral algorithm for computing these steady
states. Finally, we explore a nonexistence inequality for steady state
solutions from the recent work of Chugunova, Pugh, & Taranets.

We consider the onedimensional Maxwell equations with low contrast periodic
linear refractive index and weak Kerr nonlinearity. In this context, wave
packet initial conditions with a single carrier frequency excite infinitely
many resonances. On large but finite timescales, the coupled evolution of
backward and forward waves is governed by nonlocal equations of resonant
nonlinear geometrical optics. For the special class of solutions which are
periodic in the fast phase, these equations are equivalent to an infinite
system of nonlinear coupled mode equations, the so called it extended nonlinear
coupled mode equations, xNLCME. Numerical studies support the existence of
longlived spatially localized coherent structures, featuring a slowly varying
envelope and a train of carrier shocks.
In this paper we explore, by analytical, asymptotic and numerical methods,
the existence and properties of spatially localized structures of the xNLCME
system, which arises for a refractive index profile consisting of periodic
array of Dirac delta functions.
We consider the limit of small amplitude solutions with frequencies near a
bandedge. In this case, stationary xNLCME is wellapproximated by an infinite
system of coupled, stationary, nonlinear Schr\"odinger equations, the extended
nonlinear Schr\"odinger system, xNLS. We embed xNLS in a oneparameter family
of equations, xNLS$^\epsilon$, which interpolates between infinitely many
decoupled NLS equations ($\epsilon=0$) and xNLS ($\epsilon=1$). Using
bifurcation methods we show existence of solutions for a range of
$\epsilon\in(\epsilon_0,\epsilon_0)$ and, by a numerical continuation method,
establish the continuation of certain branches all the way to $\epsilon=1$.
Finally, we perform timedependent simulations of truncated xNLCME and find the
smallamplitude solitons to be robust to both numerical errors and the NLS
approximation.

In this article we provide numerical and analytical evidence that some
degenerate dispersive partial differential equations are illposed.
Specifically we study the K(2,2) equation $u_t = (u^2)_{xxx} + (u^2)_{x}$ and
the "degenerate Airy" equation $u_t = 2 u u_{xxx}$. For K(2,2) our results are
computational in nature: we conduct a series of numerical simulations which
demonstrate that data which is very small in $H^2$ can be of unit size at a
fixed time which is independent of the data's size. For the degenerate Airy
equation, our results are fully rigorous: we prove the existence of a compactly
supported selfsimilar solution which, when combined with certain scaling
invariances, implies illposedness (also in $H^2$).

A common challenge to proving asymptotic stability of solitary waves is
understanding the spectrum of the operator associated with the linearized flow.
The existence of eigenvalues can inhibit the dispersive estimates key to
proving stability. Following the work of Marzuola & Simpson, we prove the
absence of embedded eigenvalues for a collection of nonlinear Schrodinger
equations, including some one and three dimensional supercritical equations,
and the three dimensional cubicquintic equation. Our results also rule out
nonzero eigenvalues within the spectral gap and, in 3D, endpoint resonances.
The proof is computer assisted as it depends on the sign of certain inner
products which do not readily admit analytic representations. Our source code
is available for verification at
http://www.math.toronto.edu/simpson/files/spec_prop_asad_simpson_code.zip.

We consider the onedimensional propagation of electromagnetic waves in a
weakly nonlinear and lowcontrast spatially inhomogeneous medium with no energy
dissipation. We focus on the case of a periodic medium, in which dispersion
enters only through the (FloquetBloch) spectral band dispersion associated
with the periodic structure; chromatic dispersion (timenonlocality of the
polarization) is neglected. Numerical simulations show that for initial
conditions of wavepacket type (a plane wave of fixed carrier frequency
multiplied by a slow varying, spatially localized function) very longlived
spatially localized coherent solitonlike structures emerge, whose character is
that of a slowly varying envelope of a train of shocks. We call this structure
an envelope carriershock train.
The structure of the solution violates the oftassumed nearly monochromatic
wave packet structure, whose envelope is governed by the nonlinear coupled mode
equations (NLCME). The inconsistency and inaccuracy of NLCME lies in the
neglect of all (infinitely many) resonances except for the principle resonance
induced by the initial carrier frequency. We derive, via a nonlinear
geometrical optics expansion, a system of nonlocal integrodifferential
equations governing the coupled evolution of backward and forward propagating
waves. These equations incorporate effects of all resonances. In a periodic
medium, these equations may be expressed as a system of infinitely many coupled
mode equations, which we call the extended nonlinear coupled mode system
(xNLCME). Truncating xNLCME to include only the principle resonances leads to
the classical NLCME.
Numerical simulations of xNLCME demonstrate that it captures both large scale
features, related to third harmonic generation, and fine scale carrier shocks
features of the nonlinear periodic Maxwell equations.

The focusing cubic nonlinear Schr\"odinger equation in two dimensions admits
vortex solitons, standing wave solutions with spatial structure, Qm(r,theta) =
e^{i m theta} Rm(r). In the case of spin m = 1, we prove there exists a class
of data that collapse with the vortex soliton profile at the loglog rate. This
extends the work of Merle and Rapha\"el, (the case m = 0,) and suggests that
the L2 mass that may be concentrated at a point during generic collapse may be
unbounded. Difficulties with m >= 2 or when breaking the spin symmetry are
discussed.