Global solutions to elliptic and parabolic \({\Phi^4}\) models in Euclidean space (Q2415352): Difference between revisions

From the Introduction: ``This paper is concerned with elliptic and parabolic partial differential equations related to the \(\Phi^4\) Euclidean quantum field theory on the full space. More precisely, we consider the following semilinear elliptic PDE on \(\mathbb{R}^d\) for \(d=4,5\): \[ (-\Delta+\mu)\varphi+\varphi^3=\xi,\eqno(1.1) \] where \(\xi\) is a space white noise on \(\mathbb{R}^d\) and \(\mu>0\). We also consider the Cauchy problem for the semilinear parabolic PDE on \(\mathbb{R}_{+}\times\mathbb{R}^d\) with \(d=2,3\), given by \[ (\partial_t-\Delta+\mu)\varphi+\varphi^3=\xi,\eqno(1.2) \] where \(\xi\) is a space white noise on \(\mathbb{R}_{+}\times\mathbb{R}^d\) and \(\mu\in\mathbb{R}\). Both equations fall into the category of the so-called singular SPDEs, which means that they are classically ill-posed due to the very irregular nature of the noise \(\xi\). Indeed, solutions are expected to take values only in spaces of distributions of negative regularity and the non-linear terms appearing in the equations cannot be given a canonical meaning. [...] Our aim here is to develop a simple global solution theory for equations (1.1) and (1.2). Global solutions rely on specific properties of the equations, in particular here on the right sign of the cubic non-linearity. The existence of global in time solutions of the parabolic equation (1.2) is relevant to the problem of stochastic quantization of the \(\Phi^4\) Euclidean quantum field theory. [...] In the present work we [prove] the global space-time existence and uniqueness for equation (1.2) in \(\mathbb{R}^3\) with an associated \textit{coming down from infinity} property, meaning that, after a finite time, the solution belongs to a compact set of the state space uniformly in the initial condition. This is essentially a byproduct of the technique which we develop to analyze the elliptic model (1.1) on \(\mathbb{R}^d\) with \(d=4,5\). The choice of dimensions has a two-fold origin: first, it corresponds to the dimensions where the singularities of the elliptic equation match those of the parabolic one for \(d=2,3\). Second (and partially related reason) is that there exists a very interesting conjecture of \textit{dimensional reduction} formulated first by Parisi and Sourlas [Random magnetic fields, supersymmetry, and negative dimensions, G. Parisi, N. Sourlas, Physical Review Letters, 1979 - APS] which links the behaviour of certain SPDEs in \(d\) dimensions to that of Euclidean field theories in \(d-2\) dimensions. [...] [...] The main technical problem with globalization in the solution theory of singular PDEs is given by the fact that the noise grows at infinity, what requires the use of weighted spaces. [...] [In this work], we developed a new \textit{localization} technique which allows to split distributions belonging to [suitable] weighted spaces into an irregular component which behaves nicely at the spatial infinity and a smooth component which grows in space. The localization technique allows to split singular SPDEs into two equations: one containing the irregular terms but linear (or almost linear) and not requiring any particular care in the handling of the weighted spaces; the other containing all the more regular terms and all the non-linearities which can be analyzed using standard PDE arguments [...].''