You can always multiply two square matrices of the same size. If we multiplied a matrix by itself, that would be $M^2$. We could keep on multiplying it by itself to get arbitrarily large exponents. If you can raise matrices to powers, and multiply/divide by a real number, than can you have polynomials applied to matrices? If that is the case, then, through Taylor series, couldn't almost any function apply to square matrices? Does this have any meaningful interpretation?
-
Yes, you can do it. Here is a wikipedia link: https://en.wikipedia.org/wiki/Matrix_function I think it's used in physics as well, because we were learning about it, but I don't know where. – Botond May 21 '18 at 14:23
-
Important cases are the minimal polynomial of a matrix, that is the lowest-degree non-zero polynomial with $f(M)=0$; or the characteristic polynomial which is $\chi_M(x)=\det(xI-M)$ (and by the way is a multiple of the minimal polynomial, i.e., $\chi_M(M)=0$). The exponential function (defined over the series) has important applications; power series functions are particularly fun when fed with a nilpotent matrix as that means the power series is actually a polynomial :) – Hagen von Eitzen May 21 '18 at 14:36
2 Answers
Yes, you can apply polynomial functions to square matrices - for example, a square matrix satisfies its own characteristic polynomial.
You can also define an exponential function on square matrices, using the same Taylor series expansion as in real analysis. Proving that the Taylor series converges is a little more tricky than for real numbers, but is not too difficult. The exponential function on matrices is used in differential geometry.
Once you have the exponential function then you can define trigonometric and hyperbolic functions on square matrices too.
- 15,326
There is a whole book "Functions of Matrices: Theory and Computation" http://www.maths.manchester.ac.uk/~higham/fm/ by Nicholas J. Higham, with accompanying MATLAB Matrix Function Toolbox http://www.ma.man.ac.uk/~higham/mftoolbox/ .
The book includes a chapter on matrix logarithms. Matrix logarithms will serve you well as we boldly move into an ever more multivariate era.
- 4,496