Results 1 to 11 of 11

Math Help - Span of a vector space

  1. #1
    Member
    Joined
    Mar 2008
    Posts
    91

    Span of a vector space

    Here is a question i got for one of my exercises and I'm not too sure how i would approach such a problem.

    Is it possible to span the space of nn matrices M(n,n) using the powers
    of a single matrix A, i.e. I, A, A^2, . . ., A^n, . . .?

    First thing i thought about was if there were matrices with the property such that as u keep multiplying it by itself you would get another matrix independent of the previous ones but that turned out to be a confusing. I also tried thinking about the opposite where if all matrices eventually become linearly dependent at some point therefore disproving that there exists n*n linearly independent matrices hence proving that it is not possible to span the space. It does seem a bit complex though doing it the way i was thinking about, so any insight to how i can solve this problem simpler would be great.
    Follow Math Help Forum on Facebook and Google+

  2. #2
    Junior Member
    Joined
    Aug 2009
    Posts
    67
    taking complex matrices makes it easier to understand. any complex matrix is similar to a upper traingular one. so wlog u may assume that u started with an upper traingular one (why?).

    any power of an upper traingular matrix is also upper traingular... so all the powers will at most span only upper traingular matrices, leaving out the rest.

    for real matrices you may argue the same way using jordan forms. in this case too the answer seems to be 'no'.
    Follow Math Help Forum on Facebook and Google+

  3. #3
    Junior Member
    Joined
    Aug 2009
    Posts
    67
    looking at it another way ...

    if powers of a matrix spanned the whole space, then the smallest degree of a polynomial which is satisfied by that matrix woudld be greater than n^2 ... but you can see by the characteristic eqn that there must be a polynomial of degree n which is satisfied by that matrix.
    Follow Math Help Forum on Facebook and Google+

  4. #4
    Member
    Joined
    Mar 2008
    Posts
    91
    Quote Originally Posted by nirax View Post
    looking at it another way ...

    if powers of a matrix spanned the whole space, then the smallest degree of a polynomial which is satisfied by that matrix woudld be greater than n^2 ... but you can see by the characteristic eqn that there must be a polynomial of degree n which is satisfied by that matrix.
    can u elaborate on this please. i dont understand what u mean
    Follow Math Help Forum on Facebook and Google+

  5. #5
    ynj
    ynj is offline
    Senior Member
    Joined
    Jul 2009
    Posts
    254
    Absolutely not. If it is true, then every matrix B will be a polynomial of A. That implies that the multiplication between matrix is commutative, which is obviously impossble.
    Follow Math Help Forum on Facebook and Google+

  6. #6
    Junior Member
    Joined
    Aug 2009
    Posts
    67
    consider  A, A^2, A^3, ..., A^{n^2}, A^{{n^2}+1}, ... if this set spans the full space then you must get  {n^2} lin independent ones among these. the highest power occuring in such a set must be at least  {n^2} if you try to write an expression of the form  a_1.A + a_2.A^2 + a_3. A^3 + ... + a_i.A^{i} , this sum would never be zero unless you take i to be atleast  {n^2} by the earlier argument of lin independence.

    so the minimal polynomial of this matrix has degree at least  {n^2} , contradicting the fact that minimal poly cannot be of degree higher than  {n} .

    which step you do not follow ??
    Last edited by nirax; August 25th 2009 at 02:01 AM. Reason: clarifying
    Follow Math Help Forum on Facebook and Google+

  7. #7
    Member
    Joined
    Mar 2008
    Posts
    91
    Quote Originally Posted by nirax View Post
    consider  A, A^2, A^3, ..., A^{n^2}, A^{{n^2}+1}, ... if this set spans the full space then you must get  {n^2} lin independent ones among these. the highest power occuring in such a set must be at least  {n^2} if you try to write an expression of the form  a_1.A + a_2.A^2 + a_3. A^3 + ... + a_i.A^{i} , this sum would never be zero unless you take i to be atleast  {n^2} by the earlier argument of lin independence.

    so the minimal polynomial of this matrix has degree at least  {n^2} , contradicting the fact that minimal poly cannot be of degree higher than  {n} .

    which step you do not follow ??
    why is it the fact that minimal polynomial cannot be of degree higher than n? if u look in my OP u would see that i wrote:
    "Is it possible to span the space of nn matrices M(n,n) using the powers
    of a single matrix A, i.e. I, A, A^2, . . ., A^n, . . .?"
    so it is not limited to n
    Follow Math Help Forum on Facebook and Google+

  8. #8
    ynj
    ynj is offline
    Senior Member
    Joined
    Jul 2009
    Posts
    254
    Quote Originally Posted by ah-bee View Post
    why is it the fact that minimal polynomial cannot be of degree higher than n? if u look in my OP u would see that i wrote:
    "Is it possible to span the space of nn matrices M(n,n) using the powers
    of a single matrix A, i.e. I, A, A^2, . . ., A^n, . . .?"
    so it is not limited to n
    By the Cayley Hamilton Theorem,the eigenpolynomial of any matrix A is a annihilator of A. And any annihilator of A is a multiple of the minimal polynomial. But the degree of eigenpolynomial of A is n, which implies that the degree minimal polynimial cannot be greater than n.
    Follow Math Help Forum on Facebook and Google+

  9. #9
    Member
    Joined
    Mar 2008
    Posts
    91
    Quote Originally Posted by ynj View Post
    By the Cayley Hamilton Theorem,the eigenpolynomial of any matrix A is a annihilator of A. And any annihilator of A is a multiple of the minimal polynomial. But the degree of eigenpolynomial of A is n, which implies that the degree minimal polynimial cannot be greater than n.
    i see. thanks for that, never seen that theorem before.
    Follow Math Help Forum on Facebook and Google+

  10. #10
    ynj
    ynj is offline
    Senior Member
    Joined
    Jul 2009
    Posts
    254
    Actually, you can refer #5, which is much simpler..
    Follow Math Help Forum on Facebook and Google+

  11. #11
    Member
    Joined
    Mar 2008
    Posts
    91
    thanks.. figured it out
    Last edited by ah-bee; August 25th 2009 at 03:32 AM.
    Follow Math Help Forum on Facebook and Google+

Similar Math Help Forum Discussions

  1. Replies: 3
    Last Post: March 30th 2011, 01:19 PM
  2. How do I determine which space these vectors span?
    Posted in the Advanced Algebra Forum
    Replies: 1
    Last Post: October 4th 2010, 03:44 PM
  3. Do the 3 vectors span the space? Check :)
    Posted in the Advanced Algebra Forum
    Replies: 1
    Last Post: July 17th 2010, 04:53 AM
  4. need help...about vector space and span
    Posted in the Calculus Forum
    Replies: 4
    Last Post: September 18th 2008, 09:35 PM
  5. How to prove vectors span a space?
    Posted in the Advanced Algebra Forum
    Replies: 5
    Last Post: June 9th 2007, 02:02 PM

Search Tags


/mathhelpforum @mathhelpforum