Fractional diffusion equations and processes with randomly varying time (Q1011156)

The authors consider time-fractional diffusion equations of order \(0<\nu\leq2\) and interprete their solutions as densities of the composition of various types of stochastic processes. It is proved that the solution for \(\nu=1/2^n\) corresponds to the distribution of \(n\)-times iterated Brownian motion. The explicit solutions of the fractional diffusion equation are presented for \(\nu=1/3, 2/3, 4/3.\) In particular, \(u_{2/3}(x,t)\) is given in terms of Airy functions. The relationship between the solution \(u_{\nu}(x,t)\) and the stable densities is presented. The behavior of the solution (for \(x\) varying and \(t\) fixed), which is substantially different in the two intervals \(0<\nu\leq1\) and \(1<\nu\leq2\) is analyzed.