Certified eigenvalue location for structured matrices
Extend symbolic-numeric certified eigenvalue location to symmetric, tridiagonal and polynomial eigenvalue problems.
- Suitable for
- Master's thesis
- Supervision
- Tülay Ayyıldız
- Prerequisites
- Numerical linear algebra; some computer algebra
Certified Real Eigenvalue Location (SYNASC 2025) combines symbolic and numeric computation to locate the real eigenvalues of a matrix with guarantees. Structured matrices offer both opportunities for speed-ups and new difficulties.
Goal
Adapt the method to structured classes such as symmetric, tridiagonal and polynomial eigenvalue problems, implement it, and compare it with interval-based approaches.