Talk:Triangular matrix: Difference between revisions

Revision as of 08:59, 6 June 2014

Latest revision as of 05:31, 9 March 2024

(11 intermediate revisions by 9 users not shown)

Line 1:

{{maths rating|class=Start|importance=low|field=algebra}}

== Inverses/products of triangular matrices ==

Line 75 ⟶ 73:

In MATLAB and related programs I have seen references to 'quasi-upper-triangular' matrices, but I can't find a definition. Would someone please add a definition here? --[[User:Rinconsoleao|Rinconsoleao]] ([[User talk:Rinconsoleao|talk]]) 22:12, 28 February 2008 (UTC)

Applied Numerical Linear Algebra, James W. Demmel, 1997, copyright SIAM, page 147.

"THEOREM4.3. Real Schur canonical form. ''IF A is real, there exists a real orthogonal matrix V such that V^T A V = T is quasi-upper triangular. This means that T is block upper triangular with 1-by1 and 2-by-2 blocks on the diagonal. Its eigenvalues are the eigenvalues of the diagonal blocks. The 1-by-1 blocks correspond to real eigenvalues, and the 2-by-2 blocks to complex conjugate pairs.''

[[User:Nick Boshaft|Nick Boshaft]] ([[User talk:Nick Boshaft|talk]]) 00:48, 28 April 2016 (UTC)

== null matrix ==

Line 86 ⟶ 91:

:“Square” is generally required, square matrices being generally more interesting. For non-square matrices one generally calls these “trapezoidal” matrices, which is mentioned in the article.

:—Nils von Barth ([[User:Nbarth|nbarth]]) ([[User talk:Nbarth|talk]]) 08:24, 2 December 2009 (UTC)

::My textbook (Spence, Insel, and Friedberg. "Elementary Linear Algebra: A Matrix Approach", 2nd edition) in chapter 2.6 "The LU Decomposition of a Matrix" gives a 3x4 matrix as an example of an upper triangular matrix. Is it common to refer to non-square "trapezoidal" matrices as "triangular" in the context of LU Decompositions or is this something that is otherwise rare other than this book?

::[[User:Derek M|Derek M]] ([[User talk:Derek M|talk]]) 08:53, 17 March 2018 (UTC)

== Forward and Back Substitution ==

Line 98 ⟶ 105:

There is a lot of good material in here, but it seems to be arranged in no particular order. The level of exposition oscillates at high speed between what is appropriate for grade school and what is appropriate for graduate school. I am going to try to straighten things out a bit. Please help! [[User:Lesnail|LeSnail]] ([[User talk:Lesnail|talk]]) 01:55, 19 March 2012 (UTC)

:I've worked a bit on the first half now. The second half is untouched. [[User:Lesnail|LeSnail]] ([[User talk:Lesnail|talk]]) 03:34, 19 March 2012 (UTC)

== False claim? ==

In the article's [https://en.wikipedia.org/w/index.php?title=Triangular_matrix&oldid=761938496#Simultaneous_triangularisability section] about simultaneous triangularisability is claimed that

{{WikiProject banner shell|class=C|vital=yes|1=

{{WikiProject Mathematics|importance=mid}}

}}

{{quotation|The fact that commuting matrices have a common eigenvector can be interpreted as a result of [[Hilbert's Nullstellensatz]]: commuting matrices form a commutative algebra <math>K[A_1,\ldots,A_k]</math> over <math>K[x_1,\ldots,x_k]</math> which can be interpreted as a variety in ''k''-dimensional affine space, and '''the existence of a (common) eigenvalue''' (and hence a common eigenvector) corresponds to this variety having a point (being non-empty), which is the content of the (weak) Nullstellensatz. In algebraic terms, these operators correspond to an [[algebra representation]] of the polynomial algebra in ''k'' variables.}}

Is the claim about common eigenvalue wrong or I'm misinterpreting it? As far I know, two commuting matrices share a common eigenvector, but not necessarily a common eigenvalue: the identity matrix ''I'' and ''2I'' share common eigenvectors, but their eigenvalues are different. [[User:Saung Tadashi|Saung Tadashi]] ([[User talk:Saung Tadashi|talk]]) 23:06, 26 February 2017 (UTC)

== algorithm ==

I blanked a section labeled "Algorithm", which presented a naive block of code doing .. something. There have got to dozens if not hundreds of algorithms that can be applied to triangular matrixes. This is not the right place for a compendium of these. Readers can be referred to github for LAPACK or BLAS. [[Special:Contributions/67.198.37.16|67.198.37.16]] ([[User talk:67.198.37.16|talk]]) 17:53, 4 February 2018 (UTC)

== Chapter: Forward and back substitution ==

The algorithm for forward substitution as it is now, assumes that <math>l_{i,i}\neq 0</math>, which isn't given in general.

Furthermore, the algorithm isn't well defined for triangular matrices that don't have a staircase form, e.g.

<math>\begin{pmatrix}0&1&1\\0&1&2\\0&0&0\end{pmatrix}</math>.

I'd therefore argue that this chapter, though generally formulated for triangular matrices, would be a better fit for [[Row_echelon_form]].

This would have the added benefit that the reduced row echelon form admits a generalized backward substitution algorithm, which handles the case of underdetermined systems, and returns all possible solutions (see discussion, row echelon form)

[[User:Sanitiy|Sanitiy]] ([[User talk:Sanitiy|talk]]) 17:42, 21 December 2019 (UTC)

== Proprieties, Special Forms, Methods ==

On '''Proprieties''' it can be added:

If the inverse U−1 of an upper triangular matrix U exists, then it is upper triangular.

If the inverse L−1 of an lower triangular matrix L exists, then it is lower triangular.

http://homepages.warwick.ac.uk/~ecsgaj/matrixAlgSlidesC.pdf

Each entry on the main diagonal of L-1 is equal to the reciprocal of the corresponding entry on the main diagonal of L.

https://www.statlect.com/matrix-algebra/triangular-matrix

A Venn diagram of types of triangular matrices would be helpful in '''Special forms'''

(((Identity) Diagonal)Atomic)Uni-triangular)Non-Singular, Singular(Strictly triangular)

and bi-diagonal

Some algorithms can be added from:

"Stability of Methods for Matrix Inversion"

http://www.netlib.org/lapack/lawnspdf/lawn27.pdf

Methods:

''Unblocked method''s “i” row-wise; “j” column-wise or “k” outer products

Where i, j, k are referring to outermost loop index

For column-wise there is a method 1 and 2 discussed.

For ''blocked methods'' author presents 1B, 2B, 2C methods.

https://epubs.siam.org/doi/abs/10.1137/0119075?journalCode=smjmap [[Special:Contributions/92.120.5.12|92.120.5.12]] ([[User talk:92.120.5.12|talk]]) 11:56, 7 August 2023 (UTC)