On extension of one method for investigation of nonlinear difference schemes (Q1382864): Difference between revisions

The authors consider the following quasilinear initial-boundary value problem \[ {\partial V\over\partial t}= \text{div}(k(x,t,V)\nabla V)+ f(x,t,V,\nabla V),\quad x\in D,\quad 0<t\leq T, \] \[ V(x,0)= V^0(x),\quad V(x,t)= \mu(x,t),\quad x\in\partial D,\quad 0<t\leq T, \] where \(D\) is a bounded domain in \(\mathbb{R}^n\), \(V\) is a vector-function and the functions \(k\) and \(f\) are such that there exists a solution of the above problem. Reaction-diffusion equations, the nonlinear Schrödinger equation and the Kuramoto-Tsuzuki problem are examples of the investigated problem. The authors modify the right side of the differential equation in such a way that it does not affect the exact solution of the original problem and then they construct a standard difference scheme for the modified problem. Next, they prove (introducing suitable definitions of stability and strong stability) the correctness of the discrete problem i.e. the unique existence of the solution and the convergence to the solution of the differential problem. (If a norm in the space of mesh-functions is weaker than the maximum norm, then one has to make an additional job.) The Kuramoto-Tsuzuki problem with a nonlinear convection term is studied following the idea described above.