摘要

A hierarchical LU(H-LU) decomposition-based direct method enhanced second-order Krylov subspace projection-based model order reduction (MOR) is presented. The construction of the projection matrix is the most important process in the Krylov subspace projection-based MOR, where an inverse of the sparse matrix is involved in the construction of the projection matrix. It will be intractable for the computational capacity of a personal computer when the size of the problem is large. A H-LU decomposition-based direct method is used to the MOR with almost linear complexity. Numerical results demonstrate the validity of the proposed algorithm.

全文