2008 年 11 巻 p. 117-122
In this paper, properties of several numerical procedures based on the Galerkin method for wave equation is evaluated. Numerical properties are characterised by numerical dispersion, which is calculated by combining errors drived from both spatial and temporal discretization. The present approach enables us to understand the function of each discretization in the overall accuracy of the schemes and indicates the guideline for the design of numerical procedures.