• The Journal of Natural Sciences
• /
• v.16 no.1
• /
• pp.49-54
• /
• 2005

When teaching on the computing of eigenvalues of a matrix, we introduced the inverse power method. We compared the algorithm which is good for theoretical point of view that is useful for computational point of view.

• Communications of Mathematical Education
• /
• v.20 no.4 s.28
• /
• pp.575-586
• /
• 2006

Both theory and experiment are very important parts in sciences. Especially in mathematics, theory seems to be very important, but experiment or practice doesn't. Numerical analysis of many parts in mathematics needs practice in computer. In this paper, I suggest that computer-practicing in teaching power method, inverse power method and deflation to calculate eigenvalues and eigenvectors is good in understanding the theory. It also makes students sure that mathematics is helpful.

• The Journal of Natural Sciences
• /
• v.8 no.2
• /
• pp.9-11
• /
• 1996

We apply a full mutigrid scheme to computing eigenvalues of the Laplace eigenvalue problem with Dirichlet boundary condition. We use finite difference method to get an algebraic equation and apply inverse power method to estimating the smallest eigenvalue. Our result shows that combined method of inverse power method and full multigrid scheme is very effective in calculating eigenvalue of the eigenvalue problem.