A block-recursive approach to unitary matrix decomposition

• Published in: PROBLEMS IN PROGRAMMING
• Date: 2026
• Issue: 1
• Pages: 102-111
• ISSN: 1727-4907
• Author Affiliations: G.I. Malaschonok — National University "Kyiv-Mohyla Academy"; S.S. Sukharskyi — Institute of Software Systems NAS of Ukraine
• Keywords: унітарний розклад матриці, блоково-рекурсивний алгоритм, ортогональний розклад, сингулярний розклад, сингулярна декомпозиція матриці, паралельна програма, розріджена матриця, блочно-рекурсивні алгоритми, сингулярні числа, складність алгоритму
• Main Authors: Malaschonok, G.I., Sukharskyi, S.S.
• Format: Article
• Language: Ukrainian
• Published: PROBLEMS IN PROGRAMMING 2026
• Subjects: unitary matrix decomposition; block-recursive algorithm; parallel computing; graphics processing unit; UDC 519.61:004.272
• Online Access: https://pp.isofts.kiev.ua/index.php/ojs1/article/view/895
• Journal Title: Problems in programming

Singular value decomposition is one of the fundamental tools of numerical linear algebra and is widely used in data processing, optimization, computer vision, and machine learning. The classical approach to its computation is based on the Householder algorithm, which reduces an arbitrary matrix to a...