I have a question from my proffesor that I can not figure it out.

V will be inner product system above R2.

Let E some basis with the gram matrix (E={v1,v2}) This is the gram matrix:

\begin{pmatrix} 2 & -1\\ -1 & 1\\ \end{pmatrix} = \begin{pmatrix} {(v_{1},v_{1})} & {(v_{1},v_{2})}\\ {(v_{2},v_{1})} & {(v_{2},v_{2})}\\ \end{pmatrix}

Let T:V->V be a linear map with a matrix represent T according to basis E

\begin{pmatrix} 1 & 2\\ 2 & 1\\ \end{pmatrix}

is T is self adjoint ?

well first of all according to gram matrix it is very easy to see that E is not a orthonormal set and therefore T* is necessarily equal to the matrix of T with a transpose and conjunction.

So where can I go from here because I can not assume that the inner product is the standard one and therefore I do not know How to look for it, if exists ?

and if it is not exist how can I proof that if I do not know inner product and by that I can not find the adjoint

  • $\begingroup$ Symmetric matrices correspond to self-adjoint operators. $\endgroup$ – DVD May 16 '13 at 23:58
  • $\begingroup$ @Daved The matrix of $T$ given here corresponds to a non-orthonormal basis of $V$. So you can't read whether $T$ is self-adjoint on this matrix like that. $\endgroup$ – Julien May 17 '13 at 1:01
  • $\begingroup$ @julien okay, should add:"in orthonormal basis". $\endgroup$ – DVD May 17 '13 at 2:06
  • $\begingroup$ You should have edited your first question, where several of us took the time to respond carefully! $\endgroup$ – Ted Shifrin May 19 '13 at 0:09

Let us call $P$ the matrix of the basis $E$ expressed in an orthonormal basis, so that $$ P^*P=\pmatrix{2&-1\\-1&1}=Q $$ and, if $A$ denotes the matrix of $T$ in the same orthonormal basis, $$ P^{-1}AP=\pmatrix{1&2\\2&1}=B. $$ So the question amounts to: do we have $A^*=A$? Now, note that $B^*=B$ and $$ A^*=A\;\iff\;(PBP^{-1})^*=PBP^{-1}\;\iff\; (P^*)^{-1}BP^*=PBP^{-1}\;\iff\;BQ=QB. $$ I let you check whether the latter is true.

  • $\begingroup$ Pretty calculation!... $\endgroup$ – DVD May 17 '13 at 2:09
  • $\begingroup$ "Let us call P the matrix of the basis E expressed in an orthonormal basis, so that P∗P==Q" I . I do not understand that. How can u assume there is such a matrix P that is a matrix that transform vectors from the orthonormal representation to E such that if u multiply it by it's adjoint u will get Q. how is it possible ? besides that I understand all the rest.am I wrong thinking on the Functionality of P ? $\endgroup$ – wantToLearn May 17 '13 at 7:54
  • $\begingroup$ @julien thank u very much, but could u explain me the part I did not understand ? $\endgroup$ – wantToLearn May 17 '13 at 9:16
  • $\begingroup$ That's what a Gram matrix is. If $P=\pmatrix{a_1&b_1\\a_2&b_2}$, it means that $v_1=a_1e_1+a_2e_2$ where $(e_1,e_2)$ is the orthonormal basis we took. Then the $(1,1)$ element of $P^*P$ is $\overline{a_1}a_1+\overline{a_2}a_2=(v_1,v_1)$. Likewise the $(1,2)$ coefficient is $(v_1,v_2)$ etc... $\endgroup$ – Julien May 17 '13 at 11:59
  • $\begingroup$ @julien I will clarify my question. So I understand what P (transfer vectors from the representation according to E to the representation orthonormal basis). but how can u be sure that P satisfy the eqution P*P=Q ? $\endgroup$ – wantToLearn May 17 '13 at 15:29

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.