Solving Hard Matrix Prob: A+kB Invertible w/ Integer Entries

  • Context: Graduate 
  • Thread starter Thread starter barbiemathgurl
  • Start date Start date
  • Tags Tags
    Hard Matrix
Click For Summary

Discussion Overview

The discussion revolves around a problem involving integer matrices, specifically the conditions under which the matrix expression A+kB remains invertible for integer values of k, given that A and B are both integer matrices and certain conditions on their determinants are met. The scope includes mathematical reasoning and exploration of properties of determinants in relation to matrix invertibility.

Discussion Character

  • Mathematical reasoning
  • Exploratory
  • Debate/contested

Main Points Raised

  • One participant expresses difficulty in solving the problem and reiterates the conditions for matrices A and B, emphasizing their integer entries and invertibility.
  • Another participant suggests examining the determinants of the matrices A+kB and notes that the determinant is a polynomial in k.
  • A later reply proposes defining a function f(x) = det(A+xB) and discusses its properties, indicating that it is a polynomial of degree at most n and must be constant under certain conditions.
  • One participant mentions that the determinant being ±1 implies that the matrix is invertible with integer entries, but acknowledges a need to clarify this point further.
  • Another participant provides a hint about the polynomial f(k)^2 and its implications for the determinant, suggesting that it must be ±1 and thus reinforcing the invertibility condition.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the solution, with some providing hints and others expressing uncertainty about specific steps or assumptions. The discussion remains open-ended with multiple viewpoints on how to approach the problem.

Contextual Notes

There are unresolved assumptions regarding the properties of determinants and the implications of polynomial behavior in this context. The discussion also reflects varying levels of clarity on the triviality of certain results related to integer matrices.

barbiemathgurl
Messages
12
Reaction score
0
i just can't figure this out.

given a n x n matrix (with n>1) "A" such that all entries are integers and A is invertible such that A^{-1} also has integer entries. Let B be another matrix with integer coefficients so that:
A+B, A+2B, A+3B, ... A+(n^2)B
Are all invertible with integer entries.

Show that,
A+kB
Is also invertible with integer enties for any integer k.

who the heck do you solve this?
 
Physics news on Phys.org
What can you say about the determinants of those matrices? Once you get this, use the fact that det(A+kB), with A,B known and k a variable, is a polynomial in k.
 
barbiemathgurl said:
i just can't figure this out.

given a n x n matrix (with n>1) "A" such that all entries are integers and A is invertible such that A^{-1} also has integer entries. Let B be another matrix with integer coefficients so that:
A+B, A+2B, A+3B, ... A+(n^2)B
Are all invertible with integer entries.

Show that,
A+kB
Is also invertible with integer enties for any integer k.

who the heck do you solve this?

Let M be an arbitrary square invertible matrix whose inverse and itself has integer entires. Then [tex]1=\det (MM^{-1}) = \det(M)\det (M^{-1})[/tex] shows that [tex]\det (M) = \pm 1[/tex] because the determinant of this matrix must be an integer. Now define the function [tex]f(x) = \det (A+xB)[/tex]. This is a polyomial of at most [tex]n[/tex] degree. Notice that [tex]f(0),f(1),f(2),...,f(n^2)[/tex] are either [tex]1 \mbox{ or }-1[/tex]. By the strong form of the Pigeonhole Principle at least [tex]n+1[/tex] of them are either [tex]1[/tex] or [tex]-1[/tex]. Without lose of generality say its [tex]1[/tex]. That means [tex]f(x)[/tex] must in fact be a constant polynomial because a polynomial of at most [tex]n[/tex] degree cannot produce the same values for [tex]n+1[/tex] different values. So [tex]f(x)=1[/tex]. That means [tex]f(k)=1[/tex] for no matter what [tex]k[/tex]. So [tex]\det (A+kB)=1[/tex]. Since the determinant is 1, it must mean the matrix is irreducible with integer coefficients (by the adjoint matrix formula).
 
Kummer, we try not to give complete solutions here, just hints. And incidentally, a slightly easier way to get the last step is to note that f(k)^2 is a polynomial of degreen n^2 which is 1 at n^2+1 points, so must be 1 identically, and so f(k)=+-1. Also it remains to show that all integer matrices with determinant +-1 are invertible with integer entries.
 
StatusX said:
Kummer, we try not to give complete solutions here, just hints.
Okay.

Also it remains to show that all integer matrices with determinant +-1 are invertible with integer entries.
Last line in my first post in paranthesis. I was being sloppy on that last line because I assumed that result was trivial. I should have been more explicit.
 

Similar threads

Replies
6
Views
2K
  • · Replies 8 ·
Replies
8
Views
3K
  • · Replies 1 ·
Replies
1
Views
1K
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 17 ·
Replies
17
Views
7K
  • · Replies 2 ·
Replies
2
Views
2K
  • · Replies 2 ·
Replies
2
Views
2K
Replies
1
Views
975
  • · Replies 1 ·
Replies
1
Views
3K
  • · Replies 14 ·
Replies
14
Views
4K