Can a Matrix-Valued Function h Exist with Given Properties in GL^+(n,R)?

  • Thread starter Thread starter WWGD
  • Start date Start date
WWGD
Science Advisor
Homework Helper
Messages
7,679
Reaction score
12,446
Hi all, I am trying to construct a continuous matrix-valued function h=h(f,g) ; f,g in C^1[a,b],all defined in an open interval (a,b) on the Real line with some given properties, or at least I want to show that the function exists . I have/know that for a,b in [a,b], both Det(h(a))>0 and Det(h(b))>0 . Then I want, for all t in (a,b), that the determinant of :

A=<br /> \begin{pmatrix}<br /> f(t) &amp; g(t) \\<br /> f &#039;(t)&amp; g&#039;(t) \\<br /> <br /> \end{pmatrix}<br />

will be positive . I am trying to use that GL(n,R) has two connected components; GL^+ and
Gl^- ;matrices with positive and negative determinant respectively (e.g., use the fact that
Gaussian elimination can take any A in GL^+ into the Id. ).

Then I think can argue that since f,g ;f',g' are continuous by assumption, the map h(f,g) is
itself also continuous , so that h[a,b] -- remember here that [a,b] is an interval in the real line--is connected, and so h[a,b] must then lie in GL^+(n,R), for all t in [a,b], i.e., Det(h[a,b])>0.
Is this right?
 
Physics news on Phys.org
WWGD said:
Hi all, I am trying to construct a continuous matrix-valued function h=h(f,g) ; f,g in C^1[a,b],all defined in an open interval (a,b) on the Real line with some given properties, or at least I want to show that the function exists . I have/know that for a,b in [a,b], both Det(h(a))>0 and Det(h(b))>0 . Then I want, for all t in (a,b), that the determinant of :

A=<br /> \begin{pmatrix}<br /> f(t) &amp; g(t) \\<br /> f &#039;(t)&amp; g&#039;(t) \\<br /> <br /> \end{pmatrix}<br />

will be positive . I am trying to use that GL(n,R) has two connected components; GL^+ and
Gl^- ;matrices with positive and negative determinant respectively (e.g., use the fact that
Gaussian elimination can take any A in GL^+ into the Id. ).

Then I think can argue that since f,g ;f',g' are continuous by assumption, the map h(f,g) is
itself also continuous , so that h[a,b] -- remember here that [a,b] is an interval in the real line--is connected, and so h[a,b] must then lie in GL^+(n,R), for all t in [a,b], i.e., Det(h[a,b])>0.
Is this right?
Consider the function h:C^1([a,b])\times C^1([a,b])\to GL_1^+(\mathbb{R})\cong\mathbb{R}_+,~(f,g)\mapsto\left\{\begin{matrix}2 &amp; \text{if } f(0)+g(0)\in\mathbb{Q} \\ 1 &amp; \text{if } f(0)+g(0)\not\in\mathbb{Q}\end{matrix}\right.~.

The determinant of the 1\times 1 matrix h(f,g) is always greater than 0, and the function is nowhere continuous.

Why is A important in this? It seems unrelated.
 
Thanks, what I am actually trying to do is to start with any matrix A with DetA>0 and show it can be continuously-deformed into any other matrix B with DetB>0 , and the deformation is done within GL^+(2,R), i.e., I am trying to show GL^+(2,R) , as a Lie group (with the subspace topology of R^4)is path-connected. I think using elementary row operations--seen as continuous maps-- the right way , transforming A into the Id is /gives us the path between A and the Id, showing path-connectedness . Still, we need to find row operations that preserve the sign of Det.
I am trying to construct a collection of continuous functions parametrized by an interval , taking A to Id, i.e., a path between A and Id., to show GL^+(2,R) is path-connected (obviously this holds for GL(n,R) for all natural n) .Sorry for being unclear; I am typing with an OSK.
 
Last edited:
WWGD said:
Thanks, what I am actually trying to do is to start with any matrix A with DetA>0 and show it can be continuously-deformed into any other matrix B with DetB>0 , and the deformation is done within GL^+(2,R), i.e., I am trying to show GL^+(2,R) , as a Lie group (with the subspace topology of R^4)is path-connected. I think using elementary row operations--seen as continuous maps-- the right way , transforming A into the Id is /gives us the path between A and the Id, showing path-connectedness . Still, we need to find row operations that preserve the sign of Det.
I am trying to construct a collection of continuous functions parametrized by an interval , taking A to Id, i.e., a path between A and Id., to show GL^+(2,R) is path-connected (obviously this holds for GL(n,R) for all natural n).
When in doubt, check Stack Exchange.
 
Last edited:
Thanks, Dragon; and they also proved path-connectedness, good ref.
 
Back
Top