Maximum norm error estimates of the Crank-Nicolson scheme for solving a linear moving boundary problem (Q977441)

The authors analyze the Crank-Nicolson scheme for solving the linear convection-diffusion equation \(u_t + (au - \kappa u_x)_x = g(x,t), \quad (x,t) \in Q_T\) with the initial value condition \(u(x,0) = u_0(x), \quad x \in \Omega_0\) and the moving boundary value condition \(u(x,t) = \Phi(x,t), \quad x \in \partial \Omega_t, 0 < t \leq T\), where \(Q_T = \{(x,t) \in \mathbb{R}^2, x\in \Omega_t, t \in [0,T] \}, \Omega_t\) is an interval of \(\mathbb{R}\) for each \(t \in [0,T]\), \(a\) and \(\kappa\) are two positive constants. A linear mapping is introduced to transform the above problem to an equivalent system defined on a rectangular region. It is proved that the Crank-Nicolson scheme for the problem under consideration is unconditionally stable and convergent in the maximum norm. The convergence order is \(O(\tau^2 + h^2)\). Some numerical experiments are carried out.