Non-Gaussian limit distributions of solutions of the many-dimensional B�rgers equation with random initial data (Q4940456)

This paper deals with the multidimensional Burgers' equation. Let \(\vec u(\vec x,t)\), \(\vec x \in R^n\), \(t>0\), be a solution of the Cauchy problem in the case where the initial condition is a random field of the \(\chi^2\) type of degree \(k\) with a long-range dependence. Then the finite-dimensional distributions of the field \[ \vec X_t(\vec a)=L^{-1}(\sqrt t)t^{1/2+\alpha/2}\vec u(\vec a\sqrt t,t),\quad\vec a \in R^n,\;0<\alpha<n/2, \] converge weakly as \(t \to \infty\) to the finite-dimensional distribution of the vector random field \[ \vec X(\vec a)=C\sum_{i=0}^k\vec Y_i(\vec a),\quad\vec a \in R^n, \] where \(\vec Y_i(\vec a)\), \(i=1,\dots,k\), are independent copies of the field \[ \vec Y(\vec a)=-\int_{R^{2n}}' {{\exp\{i(\vec a,\vec \lambda_1+\vec \lambda_2)-\mu |\vec \lambda_1+\vec \lambda_2|^2\}(\vec \lambda_1+\vec \lambda_2)}\over{i|\vec \lambda_1|^{(n-\alpha)/2}|\vec \lambda_2|^{(n-\alpha)/2}}} W(d\vec \lambda_1)W(d\vec \lambda_2) \] and \(\int '\dots\) is a multiple stochastic integral with respect to a complex Gaussian white noise \(W(\cdot)\) in \(R^n.\)