Modified iterated Crank-Nicolson method with improved accuracy for advection equations.

The iterated Crank-Nicolson (ICN) method is a successful numerical algorithm for solving partial differential equations. When unequal weights are used in the ICN predictor-corrector process, the convergence rate is reduced to first order. In this paper, we propose two modified ICN algorithms that achieve second order convergence rate, based on two different ways of choosing the weights. The first approach employs geometrically averaged weights in two consecutive iterations, and the second one uses arithmetically averaged weights for two consecutive time steps. The stability and second order accuracy of our methods are verified using stability and truncation error analysis and numerically demonstrated by solving linear and semi-linear hyperbolic partial differential equations, Burgers' equation, and the Gross-Pitaevskii equation. [ABSTRACT FROM AUTHOR]