Derivation of a porous medium equation from many Markovian particles and the propagation of chaos (Q1174680)

As a model of a gas flowing through a homogeneous porous medium consider a system of \(N\) particles on the \(d\)-dimensional lattice with a discrete time parameter, which interact stochastically as follows. If there are \(m\) particles at a location \(x\), then after each unit of time each particle at \(x\) independently of the other particles jumps to one of its nearest neighbour lattice points with probability \((m/N)^{\alpha- 1}(2d)^{-1}\), or else remains at \(x\). It is shown that in a limit of many particles and continuum space and time lattice, the empirical density of the particle system converges to a weak solution \(u\) of the porous medium equation \((\partial/\partial t)u = {1\over 2}\Delta(u^ \alpha)\), \(u(0) = u_ 0\), where \(u_ 0\) is an asymptotic initial density. In addition, a propagation of chaos result is given. \textit{K. Oelschläger} [J. Differ. Equations 88, No. 2, 294-346 (1990; Zbl 0734.60101)] considered a different but related system of interacting particles in \(\mathbb{R}^ d\) evolving according to mean field gradient dynamics and obtained the same limiting density (\(\alpha = 2\)).