Homogenization and corrector theory for linear transport in random media (Q708949): Difference between revisions

This paper deals with transport equation with random attenuation and scattering coefficients \(v\cdot\nabla_x u_\varepsilon+ a_\varepsilon(x,\omega)u_\varepsilon- \int_Vk_\varepsilon(x,v',v;\omega)_\varepsilon(x,v')\,dv'=0\), \((x,v)\in X\times V\), \(u_\varepsilon(x,v)=g(x,v)\), \((x,v)\in \Gamma_-\). Here \(X\) is an open, convex, and bounded subset in \(\mathbb R^d\) for \(d=2,3\) spatial dimension, and \(V\) is the velocity space, which here is \(V=S^{d-1}\) -- the unit sphere; \(\Gamma_\pm=\{(x,v)\mid x\in \partial X,\pm\nu_x\cdot v>0\}\), where \(\partial X\) is the boundary of \(X\), assumed to be smooth, and the normal vector to \(X\) at \(x\in\partial X\) is denoted by \(\nu_x\). The authors show, as \(\varepsilon\downarrow 0\), that the heterogeneous transport solution is well-approximated by a homogeneous transport solution. Then it is shown that the rescaled corrector converges in probability distribution and weakly in the space and velocity variables to a Gaussian process as an application of the central limit theorem.