Degenerate diffusion with a drift potential: a viscosity solutions approach (Q972052)

This paper concerns the pressure equation \[ u_t = (m-1)u\Delta u + |\nabla u|^2 +\nabla u\cdot \nabla \Phi + (m-1) u\Delta \Phi \] for the porous medium equation with a drift potential \(\Phi\): \[ \rho_t = \Delta(\rho^m) + \nabla\cdot(\rho\nabla\Phi), \] where \(m > 1,\;u = \frac m{m-1}\rho^{m-1}\), and \(\Phi : \mathbb R^n \to \mathbb R\) is a \(C^2\) function. The initial condition \(u(x,0) = u_0(x)\) (or \(\rho(x,0) = \rho_0(x)\)) is imposed, where \(u_0(x)\) (or \(\rho_0(x)\)) is a nonnegative continuous function having compact support. The pressure equation with \(\Phi = |x|^2\) is obtained as a re-scaled version of the pressure equation with \(\Phi \equiv 0\) by introducing new variables \[ \theta(\eta,\tau) = t^{-\alpha}u(x,t),\quad \eta = xt^{-\beta},\quad \tau = \log t, \] where \(u\) solves the pressure equation with \(\Phi \equiv 0\). This suggests that the local behavior of the solution of the pressure equation with \(\Phi\) is similar to that of the equation with \(\Phi \equiv 0\). The notion of viscosity solutions of the pressure equation with \(\Phi \equiv 0\) has been introduced by \textit{L. Caffarelli} and \textit{J. L. Vazquez} [Proc. Symp. Pure Math. 65, 13--26 (1999; Zbl 0929.35072)]. Here the authors introduce a notion of viscosity solutions of the pressure equation with general \(\Phi\) to show that these viscosity solutions coincide with the weak solutions in the distribution sense. The use of the viscosity solutions enables them to describe the pointwise behavior of the free boundary evolution by maximum principle arguments with local barriers. It is shown that, if \(|\nabla \Phi| > 0\) except at \(x = x_0\), where \(\Phi\) achieves its minimum, then the free boundary uniformly converges to the equilibrium as \(t \to \infty\). Moreover, when \(\Phi\) is strictly convex, an exponential rate of the convergence is obtained.