
The randomfield Ising model (RFIM) is one of the simplest
statisticalmechanical models that captures the anomalous irreversible
collective response seen in a wide range of physical, biological, or
socioeconomic situations in the presence of interactions and intrinsic
heterogeneity or disorder. When slowly driven at zero temperature it can
display an outofequilibrium phase transition associated with critical scaling
("crackling noise"), while it undergoes at equilibrium, under either
temperature or disorderstrength changes, a thermodynamic phase transition. We
show that the outofequilibrium and equilibrium critical behaviors are in the
same universality class: they are controlled, in the renormalizationgroup (RG)
sense, by the same zerotemperature fixed point. We do so by combining a
fieldtheoretical formalism that accounts for the multiple metastable states
and the exact (functional) RG. As a spinoff, we also demonstrate that critical
fluids in disordered porous media are in the same universality class as the
RFIM, thereby unifying a broad spectrum of equilibrium and outofequilibrium
phenomena.

We show that, contrary to previous suggestions based on computer simulations
or erroneous theoretical treatments, the critical points of the randomfield
Ising model out of equilibrium, when quasistatically changing the applied
source at zero temperature, and in equilibrium are not in the same universality
class below some critical dimension $d_{DR}\approx 5.1$. We demonstrate this by
implementing a nonperturbative functional renormalization group for the
associated dynamical field theory. Above $d_{DR}$, the avalanches, which
characterize the evolution of the system at zero temperature, become irrelevant
at large distance, and hysteresis and equilibrium critical points are then
controlled by the same fixed point. We explain how to use computer simulation
and finitesize scaling to check the correspondence between in and out of
equilibrium criticality in a far less ambiguous way than done so far.

We present a numerical study of the manybody localization (MBL) phenomenon
in the hightemperature limit within an anisotropic Heisenberg model with
random local fields. Taking the dynamical spin conductivity $\sigma(\omega)$ as
the test quantity, we investigate the full frequency dependence of
sampletosample fluctuations and their scaling properties as a function of the
system size $L\leq 28$ and the frequency resolution. We identify differences
between the general interacting case $\Delta>0$ and the anisotropy $\Delta=0$,
the latter corresponding to the standard Anderson localization. Except for the
extreme MBL case when the relative sampletosample fluctuations became large,
numerical results allow for the extraction of the low$\omega$ dependence of
the conductivity. Results for the d.c. value $\sigma_0$ indicate a crossover
into the MBL regime, i.e. an exponentiallike variation with the disorder
strength $W$. For the same regime, our numerical analysis indicates that the
lowfrequency exponent $\alpha$ exhibits a small departure from $\alpha\sim 1$
only.

The randomfield Ising model shows extreme critical slowdown that has been
described by activated dynamic scaling: the characteristic time for the
relaxation to equilibrium diverges exponentially with the correlation length,
$\ln \tau\sim \xi^\psi/T$ , with $\psi$ an \textit{a priori} unknown barrier
exponent. Through a nonperturbative functional renormalization group, we show
that for spatial dimensions $d$ less than a critical value $d_{DR} \simeq 5.1$,
also associated with dimensionalreduction breakdown, $\psi=\theta$ with
$\theta$ the temperature exponent near the zerotemperature fixed point that
controls the critical behavior. For $d>d_{DR}$ on the other hand,
$\psi=\theta2\lambda$ where $\theta=2$ and $\lambda>0$ a new exponent. At the
upper critical dimension $d=6$, $\lambda=1$ so that $\psi=0$, and activated
scaling gives way to conventional scaling. We give a physical interpretation of
the results in terms of collective events in real space, avalanches and
droplets. We also propose a way to check the two regimes by computer
simulations of longrange 1$d$ systems.

We study the critical behavior of the onedimensional random field Ising
model (RFIM) with longrange interactions ($\propto r^{(d+\sigma)}$) by the
nonperturbative functional renormalization group. We find two distinct regimes
of critical behavior as a function of $\sigma$, separated by a critical value
$\sigma_c$. What distinguishes these two regimes is the presence or not of a
cusplike nonanalyticity in the functional dependence of the renormalized
cumulants of the random field at the fixed point. This change of behavior can
be associated to the characteristics of the largescale avalanches present in
the system at zero temperature. We propose ways to check these predictions
through lattice simulations. We also discuss the difference with the RFIM on
the Dyson hierarchical lattice.

We consider the zerotemperature fixed points controlling the critical
behavior of the $d$dimensional randomfield Ising, and more generally $O(N)$,
models. We clarify the nature of these fixed points and their stability in the
region of the $(N,d)$ plane where one passes from a critical behavior
satisfying the $d\rightarrow d2$ dimensional reduction to one where it breaks
down due to the appearance of strong enough nonanalyticities in the functional
dependence of the cumulants of the renormalized disorder. We unveil an
intricate and unusual behavior.

We show that the critical scaling behavior of randomfield systems with
shortrange interactions and disorder correlations cannot be described in
general by only two independent exponents, contrary to previous claims. This
conclusion is based on a theoretical description of the whole (d,N) domain of
the ddimensional randomfield O(N) model and points to the role of rare events
that are overlooked by the proposed derivations of twoexponent scaling. Quite
strikingly, however, the numerical estimates of the critical exponents of the
random field Ising model are extremely close to the predictions of the
twoexponent scaling, so that the issue cannot be decided on the basis of
numerical simulations.

We have done a finitesize scaling study of a continuous phase transition
altered by the quenched bond disorder, investigating systems at quasicritical
temperatures of each disorder realization by using the equilibriumlike invaded
cluster algorithm. Our results indicate that in order to access the thermal
critical exponent $y_\tau$, it is necessary to average the free energy at
quasicritical temperatures of each disorder configuration. Despite the thermal
fluctuations on the scale of the system at the transition point, we find that
spatial inhomogeneities form in the system and become more pronounced as the
size of the system increases. This leads to different exponents describing
rescaling of the fluctuations of observables in disorder and thermodynamic
ensembles.

We present a detailed study of the Equilibriumlike invaded cluster algorithm
(EIC), recently proposed as an extension of the invaded cluster (IC) algorithm,
designed to drive the system to criticality while still preserving the
equilibrium ensemble. We perform extensive simulations on two special cases of
the Potts model and examine the precision of critical exponents by including
the leading corrections. We show that both thermal and magnetic critical
exponents can be obtained with high accuracy compared to the best available
results. The choice of the auxiliary parameters of the algorithm is discussed
in context of dynamical properties. We also discuss the relation to the
LiSokal bound for the dynamical exponent $z$.

The invaded cluster approach is extended to 2D Potts model with annealed
vacancies by using the randomcluster representation. Geometrical arguments are
used to propose the algorithm which converges to the tricritical point in the
twodimensional parameter space spanned by temperature and the chemical
potential of vacancies. The tricritical point is identified as a simultaneous
onset of the percolation of a FortuinKasteleyn cluster and of a percolation of
"geometrical disorder cluster". The location of the tricritical point and the
concentration of vacancies for q = 1, 2, 3 are found to be in good agreement
with the best known results. Scaling properties of the percolating scaling
cluster and related critical exponents are also presented.