Monotone method for diffusion equations with nonlinear diffusion coefficients

From MaRDI portal





Consider the following quasilinear equation \[ u_t-\sum_{i, j}\bigl(a^{i,j} (x,t,u)u_i \bigr)_j= f(x,t,u,Du) \tag{1} \] for parabolic problems, and \[ -\sum_{i,j} \bigl(a^{i,j} (x,u)u_i\bigr)_j =f(x,u,Du)\tag{2} \] for elliptic problems. Denote the left-hand side of equation (1) or equation (2) by \({\mathcal A}(u)u\). Unfortunately, when trying to use the traditional monotone scheme \[ {\mathcal A}(u^{(k-1)}) u^{(k)}+ b^{(k)}u^{(k)}= b^{(k-1)}+ F(\cdot, u^{(k)}, Du^{(k)}) \tag{3} \] to equation (1) or equation (2), one faces the difficulty in determining the iterative coefficient \(b^{(k)}\) such that the sequence \(\{u^{(k)}\}\) is monotone and its limit is a solution of equation (1) or equation (2). The point is that \(b^{(k)}\) depends not only on \(u^{(k)}\) as the traditional case but also on the derivatives up to second order of \(u^{(k)}\) which are unknown in the current iterative process. In this paper, the author develops a scheme for the choice of \(b^{(k)}\) such that 1. the sequence \(\{u^{(k)}\}\) constructed from equation (3) with suitable boundary (and initial) conditions exists and is monotone, and 2. \(\{u^{(k)}\}\) converges to a solution of equation (1) or equation (2) as \(k\to\infty\).




Cited in
(23)








This page was built for publication: Monotone method for diffusion equations with nonlinear diffusion coefficients

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4259714)