We consider the perturbation formula for the weighted Moore-Penrose inverse of a rectangular matrix and give an explicit expression for the weighted Moore-Penrose inverse of a perturbed matrix under the weakest rank condition. This explicit expression extends the earlier work of several authors. (~)
Weighted Moore-Penrose inverse of a boolean matrix
✍ Scribed by R.B. Bapat; S.K. Jain; S. Pati
- Publisher
- Elsevier Science
- Year
- 1997
- Tongue
- English
- Weight
- 609 KB
- Volume
- 255
- Category
- Article
- ISSN
- 0024-3795
No coin nor oath required. For personal study only.
✦ Synopsis
If A is a boolean matrix, then the weighted Moore-Penrose inverse of A (with respect to the given matrices M, N) IS a matrix G which satisfies AGA = A, GAG = G, and that MAG and GAN are symmetric. Under certain conditions on M, N it is shown that the weighted Moore-Penrose inverse exists if and only if ANATMA = A, in which case the inverse is NTATM T. When M, N are identity matrices, this reduces to the well-known result that the Moore-Penrose inverse of a boolean matrix, when it exists, must be AT. 0
📜 SIMILAR VOLUMES
Consider an arbitrary symmetric nonnegative de®nite matrix A and its Moore± Penrose inverse A , partitioned, respectively as
In this paper, we consider the product of matrices P AQ, where A is von Neumann regular and there exist P and Q such that P P A = A = AQQ . We give necessary and sufficient conditions in order to P AQ be Moore-Penrose invertible, extending known characterizations. Finally, an application is given to