Optimal smoothing in function-transport particle methods for diffusion problems (Q1322999)

The authors consider the solution of the diffusion problem \({\partial v\over\partial t}=\nu{\partial^ 2v\over\partial x^ 2}\), \(v(x,0)=f(x)\), where the diffusion coefficient \(\nu\) is small. They discuss the modelling of diffusion by a random walk update procedure of particle positions. They give conditions under which this process can lead to a valid approximate solution which converges uniformly in \(x\). They illustrate the applicability of the method for two particular forms of initial data and discuss, in particular the case for \(\nu\) very small.