Conditions for a certain parametric matrix to be positive definite











up vote
1
down vote

favorite












The task:




Let's consider matrix $A=begin{bmatrix}1&a&a\a&1&a\a&a&1 end{bmatrix}$. Show that $A>0$ if and only if $-1<2a<2$.




Solution attempt: By definition $A$ is positive definite iff $forall_{zinmathbb{R}^3}z^{operatorname{T}}Az>0$.



For $z=begin{bmatrix}z_1\z_2\z_3end{bmatrix}$, $$left(z^{operatorname{T}}Aright)z=left(z_1^2+az_1z_2+az_1z_3right)+left(az_1z_2+z_2^2+az_2z_3right)+left(az_1z_3+az_2z_3+z_3^2right)\=z_1^2+z_2^2+z_3^2+2aleft(z_1z_2+z_1z_3+z_2z_3right)>0\iff\z_1^2+z_2^2+z_3^2>-2aleft(z_1z_2+z_1z_3+z_2z_3right)$$



This is trivially true for $z_1z_2+z_1z_3+z_2z_3geq0$, so without loss of generality let's assume that $z_1z_2+z_1z_3+z_2z_3<0$, and thus $-(z_1z_2+z_1z_3+z_2z_3)>0$. Therefore we need to prove that:



$$left({Largeforall_{z_1,z_2,z_3}}-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}>2aright)iff -1<2a<2$$



And this seems false to me. This is because $-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}$ is always $geq0$. For this reason we have a trivial counterexample, eg when $2a=-1000$ then $2anotin(-1;2)$ but nonetheless $forall_{z_1,z_2,z_3}-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}>2a$.



I can't find my error; could you kindly help me please?










share|cite|improve this question




















  • 1




    Do you know what the chracteristic polynomial of a matrix is? Its roots are the eigenvalues. In this case, one can compute the eigenvalues of the given matrix explicitly using this approach. A matrix is positive definite if and only if all its eigenvalues are positive and real.
    – астон вілла олоф мэллбэрг
    Nov 13 at 17:10












  • @астонвіллаолофмэллбэрг I know what the characteristic polynomial is, but I don't know what is the correlation between eigenvalues and positive definiteness...
    – gaazkam
    Nov 13 at 17:13






  • 1




    Your error is when you say "this is trivially true for $z_1z_2+z_1z_3+z_2z_3 geq 0$ ": that's not true for $a<0$
    – krirkrirk
    Nov 13 at 17:14










  • @астонвіллаолофмэллбэрг OK: I checked Wikipedia: All eigenvalues are positive iff matrix is positive definite. Thank you, I'll try to solve it that way
    – gaazkam
    Nov 13 at 17:20










  • Yes, I think the CP is the easiest way of doing this.
    – астон вілла олоф мэллбэрг
    Nov 13 at 17:23















up vote
1
down vote

favorite












The task:




Let's consider matrix $A=begin{bmatrix}1&a&a\a&1&a\a&a&1 end{bmatrix}$. Show that $A>0$ if and only if $-1<2a<2$.




Solution attempt: By definition $A$ is positive definite iff $forall_{zinmathbb{R}^3}z^{operatorname{T}}Az>0$.



For $z=begin{bmatrix}z_1\z_2\z_3end{bmatrix}$, $$left(z^{operatorname{T}}Aright)z=left(z_1^2+az_1z_2+az_1z_3right)+left(az_1z_2+z_2^2+az_2z_3right)+left(az_1z_3+az_2z_3+z_3^2right)\=z_1^2+z_2^2+z_3^2+2aleft(z_1z_2+z_1z_3+z_2z_3right)>0\iff\z_1^2+z_2^2+z_3^2>-2aleft(z_1z_2+z_1z_3+z_2z_3right)$$



This is trivially true for $z_1z_2+z_1z_3+z_2z_3geq0$, so without loss of generality let's assume that $z_1z_2+z_1z_3+z_2z_3<0$, and thus $-(z_1z_2+z_1z_3+z_2z_3)>0$. Therefore we need to prove that:



$$left({Largeforall_{z_1,z_2,z_3}}-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}>2aright)iff -1<2a<2$$



And this seems false to me. This is because $-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}$ is always $geq0$. For this reason we have a trivial counterexample, eg when $2a=-1000$ then $2anotin(-1;2)$ but nonetheless $forall_{z_1,z_2,z_3}-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}>2a$.



I can't find my error; could you kindly help me please?










share|cite|improve this question




















  • 1




    Do you know what the chracteristic polynomial of a matrix is? Its roots are the eigenvalues. In this case, one can compute the eigenvalues of the given matrix explicitly using this approach. A matrix is positive definite if and only if all its eigenvalues are positive and real.
    – астон вілла олоф мэллбэрг
    Nov 13 at 17:10












  • @астонвіллаолофмэллбэрг I know what the characteristic polynomial is, but I don't know what is the correlation between eigenvalues and positive definiteness...
    – gaazkam
    Nov 13 at 17:13






  • 1




    Your error is when you say "this is trivially true for $z_1z_2+z_1z_3+z_2z_3 geq 0$ ": that's not true for $a<0$
    – krirkrirk
    Nov 13 at 17:14










  • @астонвіллаолофмэллбэрг OK: I checked Wikipedia: All eigenvalues are positive iff matrix is positive definite. Thank you, I'll try to solve it that way
    – gaazkam
    Nov 13 at 17:20










  • Yes, I think the CP is the easiest way of doing this.
    – астон вілла олоф мэллбэрг
    Nov 13 at 17:23













up vote
1
down vote

favorite









up vote
1
down vote

favorite











The task:




Let's consider matrix $A=begin{bmatrix}1&a&a\a&1&a\a&a&1 end{bmatrix}$. Show that $A>0$ if and only if $-1<2a<2$.




Solution attempt: By definition $A$ is positive definite iff $forall_{zinmathbb{R}^3}z^{operatorname{T}}Az>0$.



For $z=begin{bmatrix}z_1\z_2\z_3end{bmatrix}$, $$left(z^{operatorname{T}}Aright)z=left(z_1^2+az_1z_2+az_1z_3right)+left(az_1z_2+z_2^2+az_2z_3right)+left(az_1z_3+az_2z_3+z_3^2right)\=z_1^2+z_2^2+z_3^2+2aleft(z_1z_2+z_1z_3+z_2z_3right)>0\iff\z_1^2+z_2^2+z_3^2>-2aleft(z_1z_2+z_1z_3+z_2z_3right)$$



This is trivially true for $z_1z_2+z_1z_3+z_2z_3geq0$, so without loss of generality let's assume that $z_1z_2+z_1z_3+z_2z_3<0$, and thus $-(z_1z_2+z_1z_3+z_2z_3)>0$. Therefore we need to prove that:



$$left({Largeforall_{z_1,z_2,z_3}}-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}>2aright)iff -1<2a<2$$



And this seems false to me. This is because $-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}$ is always $geq0$. For this reason we have a trivial counterexample, eg when $2a=-1000$ then $2anotin(-1;2)$ but nonetheless $forall_{z_1,z_2,z_3}-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}>2a$.



I can't find my error; could you kindly help me please?










share|cite|improve this question















The task:




Let's consider matrix $A=begin{bmatrix}1&a&a\a&1&a\a&a&1 end{bmatrix}$. Show that $A>0$ if and only if $-1<2a<2$.




Solution attempt: By definition $A$ is positive definite iff $forall_{zinmathbb{R}^3}z^{operatorname{T}}Az>0$.



For $z=begin{bmatrix}z_1\z_2\z_3end{bmatrix}$, $$left(z^{operatorname{T}}Aright)z=left(z_1^2+az_1z_2+az_1z_3right)+left(az_1z_2+z_2^2+az_2z_3right)+left(az_1z_3+az_2z_3+z_3^2right)\=z_1^2+z_2^2+z_3^2+2aleft(z_1z_2+z_1z_3+z_2z_3right)>0\iff\z_1^2+z_2^2+z_3^2>-2aleft(z_1z_2+z_1z_3+z_2z_3right)$$



This is trivially true for $z_1z_2+z_1z_3+z_2z_3geq0$, so without loss of generality let's assume that $z_1z_2+z_1z_3+z_2z_3<0$, and thus $-(z_1z_2+z_1z_3+z_2z_3)>0$. Therefore we need to prove that:



$$left({Largeforall_{z_1,z_2,z_3}}-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}>2aright)iff -1<2a<2$$



And this seems false to me. This is because $-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}$ is always $geq0$. For this reason we have a trivial counterexample, eg when $2a=-1000$ then $2anotin(-1;2)$ but nonetheless $forall_{z_1,z_2,z_3}-frac{z_1^2+z_2^2+z_3^2}{z_1z_2+z_1z_3+z_2z_3}>2a$.



I can't find my error; could you kindly help me please?







vectors matrix-equations positive-definite






share|cite|improve this question















share|cite|improve this question













share|cite|improve this question




share|cite|improve this question








edited Nov 13 at 16:54









Henning Makholm

235k16299534




235k16299534










asked Nov 13 at 16:52









gaazkam

418313




418313








  • 1




    Do you know what the chracteristic polynomial of a matrix is? Its roots are the eigenvalues. In this case, one can compute the eigenvalues of the given matrix explicitly using this approach. A matrix is positive definite if and only if all its eigenvalues are positive and real.
    – астон вілла олоф мэллбэрг
    Nov 13 at 17:10












  • @астонвіллаолофмэллбэрг I know what the characteristic polynomial is, but I don't know what is the correlation between eigenvalues and positive definiteness...
    – gaazkam
    Nov 13 at 17:13






  • 1




    Your error is when you say "this is trivially true for $z_1z_2+z_1z_3+z_2z_3 geq 0$ ": that's not true for $a<0$
    – krirkrirk
    Nov 13 at 17:14










  • @астонвіллаолофмэллбэрг OK: I checked Wikipedia: All eigenvalues are positive iff matrix is positive definite. Thank you, I'll try to solve it that way
    – gaazkam
    Nov 13 at 17:20










  • Yes, I think the CP is the easiest way of doing this.
    – астон вілла олоф мэллбэрг
    Nov 13 at 17:23














  • 1




    Do you know what the chracteristic polynomial of a matrix is? Its roots are the eigenvalues. In this case, one can compute the eigenvalues of the given matrix explicitly using this approach. A matrix is positive definite if and only if all its eigenvalues are positive and real.
    – астон вілла олоф мэллбэрг
    Nov 13 at 17:10












  • @астонвіллаолофмэллбэрг I know what the characteristic polynomial is, but I don't know what is the correlation between eigenvalues and positive definiteness...
    – gaazkam
    Nov 13 at 17:13






  • 1




    Your error is when you say "this is trivially true for $z_1z_2+z_1z_3+z_2z_3 geq 0$ ": that's not true for $a<0$
    – krirkrirk
    Nov 13 at 17:14










  • @астонвіллаолофмэллбэрг OK: I checked Wikipedia: All eigenvalues are positive iff matrix is positive definite. Thank you, I'll try to solve it that way
    – gaazkam
    Nov 13 at 17:20










  • Yes, I think the CP is the easiest way of doing this.
    – астон вілла олоф мэллбэрг
    Nov 13 at 17:23








1




1




Do you know what the chracteristic polynomial of a matrix is? Its roots are the eigenvalues. In this case, one can compute the eigenvalues of the given matrix explicitly using this approach. A matrix is positive definite if and only if all its eigenvalues are positive and real.
– астон вілла олоф мэллбэрг
Nov 13 at 17:10






Do you know what the chracteristic polynomial of a matrix is? Its roots are the eigenvalues. In this case, one can compute the eigenvalues of the given matrix explicitly using this approach. A matrix is positive definite if and only if all its eigenvalues are positive and real.
– астон вілла олоф мэллбэрг
Nov 13 at 17:10














@астонвіллаолофмэллбэрг I know what the characteristic polynomial is, but I don't know what is the correlation between eigenvalues and positive definiteness...
– gaazkam
Nov 13 at 17:13




@астонвіллаолофмэллбэрг I know what the characteristic polynomial is, but I don't know what is the correlation between eigenvalues and positive definiteness...
– gaazkam
Nov 13 at 17:13




1




1




Your error is when you say "this is trivially true for $z_1z_2+z_1z_3+z_2z_3 geq 0$ ": that's not true for $a<0$
– krirkrirk
Nov 13 at 17:14




Your error is when you say "this is trivially true for $z_1z_2+z_1z_3+z_2z_3 geq 0$ ": that's not true for $a<0$
– krirkrirk
Nov 13 at 17:14












@астонвіллаолофмэллбэрг OK: I checked Wikipedia: All eigenvalues are positive iff matrix is positive definite. Thank you, I'll try to solve it that way
– gaazkam
Nov 13 at 17:20




@астонвіллаолофмэллбэрг OK: I checked Wikipedia: All eigenvalues are positive iff matrix is positive definite. Thank you, I'll try to solve it that way
– gaazkam
Nov 13 at 17:20












Yes, I think the CP is the easiest way of doing this.
– астон вілла олоф мэллбэрг
Nov 13 at 17:23




Yes, I think the CP is the easiest way of doing this.
– астон вілла олоф мэллбэрг
Nov 13 at 17:23










2 Answers
2






active

oldest

votes

















up vote
2
down vote














$z_1^2+z_2^2+z_3^2+2aleft(z_1z_2+z_1z_3+z_2z_3right)>0iff z_1^2+z_2^2+z_3^2>-2aleft(z_1z_2+z_1z_3+z_2z_3right)$



This is trivially true for $z_1z_2+z_1z_3+z_2z_3geq0$




Not, it is not, since $a$ might be negative.



I suppose showing $A$ is definite positive could be done your way, but it's way too exhausting for me. I'd go with using the fact that $$Atext{ definite positive } iff text{ its eigenvalues are } > 0$$



Now choose your favorite way to find the eigenvalues of $A$. Note that the sum of the rows are $2a+1$ so this is one of them ; which by the way leads to $-1<2a$. Now your turn...






share|cite|improve this answer




























    up vote
    1
    down vote













    Let's determine the eigenvalues of $A$. If we can find the conditions for which those eigenvalues are all positive, then we can solve the question.



    Let $u in mathbb R^3$ be the vector with $1$ for all its components. Then note that $uu^T$ is the $3times 3$ matrix with $1$ for all its entries.
    Therefore $$A=(1-a)I+auu^T$$
    Now, notice that $$Au=(1-a)u+au (u^Tu)=(1-a)u+3au=(1+2a)u$$
    So $u$ is an eigenvector of $A$ with eigenvalue $(1+2a)$.



    Now, let's determine the other eigenvalues of $A$. Suppose $xinmathbb R^3$ is orthogonal to $u$, that is $u^Tx=0$, then
    $$Ax=(1-a)x+auu^Tx=(1-a)x$$
    so $x$ is an eigenvector of $A$ with eigenvalue $(1-a)$. Since the space of vectors $x$ orthogonal to $u$ is of dimension 2, we have now determined that $(1-a)$ is an eigenvalue of multiplicity 2 for $A$.



    In summary, the spectrum of $A$ is $(1+2a)$ (multiplicity 1) and $(1-a)$ (multiplicity 2). Consequently, $A$ is positive definite iff $$1+2a>0 text{ and } 1-a >0$$ that is iff $-1 < 2a < 2$$






    share|cite|improve this answer





















      Your Answer





      StackExchange.ifUsing("editor", function () {
      return StackExchange.using("mathjaxEditing", function () {
      StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
      StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
      });
      });
      }, "mathjax-editing");

      StackExchange.ready(function() {
      var channelOptions = {
      tags: "".split(" "),
      id: "69"
      };
      initTagRenderer("".split(" "), "".split(" "), channelOptions);

      StackExchange.using("externalEditor", function() {
      // Have to fire editor after snippets, if snippets enabled
      if (StackExchange.settings.snippets.snippetsEnabled) {
      StackExchange.using("snippets", function() {
      createEditor();
      });
      }
      else {
      createEditor();
      }
      });

      function createEditor() {
      StackExchange.prepareEditor({
      heartbeatType: 'answer',
      convertImagesToLinks: true,
      noModals: true,
      showLowRepImageUploadWarning: true,
      reputationToPostImages: 10,
      bindNavPrevention: true,
      postfix: "",
      imageUploader: {
      brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
      contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
      allowUrls: true
      },
      noCode: true, onDemand: true,
      discardSelector: ".discard-answer"
      ,immediatelyShowMarkdownHelp:true
      });


      }
      });














       

      draft saved


      draft discarded


















      StackExchange.ready(
      function () {
      StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2996981%2fconditions-for-a-certain-parametric-matrix-to-be-positive-definite%23new-answer', 'question_page');
      }
      );

      Post as a guest















      Required, but never shown

























      2 Answers
      2






      active

      oldest

      votes








      2 Answers
      2






      active

      oldest

      votes









      active

      oldest

      votes






      active

      oldest

      votes








      up vote
      2
      down vote














      $z_1^2+z_2^2+z_3^2+2aleft(z_1z_2+z_1z_3+z_2z_3right)>0iff z_1^2+z_2^2+z_3^2>-2aleft(z_1z_2+z_1z_3+z_2z_3right)$



      This is trivially true for $z_1z_2+z_1z_3+z_2z_3geq0$




      Not, it is not, since $a$ might be negative.



      I suppose showing $A$ is definite positive could be done your way, but it's way too exhausting for me. I'd go with using the fact that $$Atext{ definite positive } iff text{ its eigenvalues are } > 0$$



      Now choose your favorite way to find the eigenvalues of $A$. Note that the sum of the rows are $2a+1$ so this is one of them ; which by the way leads to $-1<2a$. Now your turn...






      share|cite|improve this answer

























        up vote
        2
        down vote














        $z_1^2+z_2^2+z_3^2+2aleft(z_1z_2+z_1z_3+z_2z_3right)>0iff z_1^2+z_2^2+z_3^2>-2aleft(z_1z_2+z_1z_3+z_2z_3right)$



        This is trivially true for $z_1z_2+z_1z_3+z_2z_3geq0$




        Not, it is not, since $a$ might be negative.



        I suppose showing $A$ is definite positive could be done your way, but it's way too exhausting for me. I'd go with using the fact that $$Atext{ definite positive } iff text{ its eigenvalues are } > 0$$



        Now choose your favorite way to find the eigenvalues of $A$. Note that the sum of the rows are $2a+1$ so this is one of them ; which by the way leads to $-1<2a$. Now your turn...






        share|cite|improve this answer























          up vote
          2
          down vote










          up vote
          2
          down vote










          $z_1^2+z_2^2+z_3^2+2aleft(z_1z_2+z_1z_3+z_2z_3right)>0iff z_1^2+z_2^2+z_3^2>-2aleft(z_1z_2+z_1z_3+z_2z_3right)$



          This is trivially true for $z_1z_2+z_1z_3+z_2z_3geq0$




          Not, it is not, since $a$ might be negative.



          I suppose showing $A$ is definite positive could be done your way, but it's way too exhausting for me. I'd go with using the fact that $$Atext{ definite positive } iff text{ its eigenvalues are } > 0$$



          Now choose your favorite way to find the eigenvalues of $A$. Note that the sum of the rows are $2a+1$ so this is one of them ; which by the way leads to $-1<2a$. Now your turn...






          share|cite|improve this answer













          $z_1^2+z_2^2+z_3^2+2aleft(z_1z_2+z_1z_3+z_2z_3right)>0iff z_1^2+z_2^2+z_3^2>-2aleft(z_1z_2+z_1z_3+z_2z_3right)$



          This is trivially true for $z_1z_2+z_1z_3+z_2z_3geq0$




          Not, it is not, since $a$ might be negative.



          I suppose showing $A$ is definite positive could be done your way, but it's way too exhausting for me. I'd go with using the fact that $$Atext{ definite positive } iff text{ its eigenvalues are } > 0$$



          Now choose your favorite way to find the eigenvalues of $A$. Note that the sum of the rows are $2a+1$ so this is one of them ; which by the way leads to $-1<2a$. Now your turn...







          share|cite|improve this answer












          share|cite|improve this answer



          share|cite|improve this answer










          answered Nov 13 at 17:20









          krirkrirk

          1,462518




          1,462518






















              up vote
              1
              down vote













              Let's determine the eigenvalues of $A$. If we can find the conditions for which those eigenvalues are all positive, then we can solve the question.



              Let $u in mathbb R^3$ be the vector with $1$ for all its components. Then note that $uu^T$ is the $3times 3$ matrix with $1$ for all its entries.
              Therefore $$A=(1-a)I+auu^T$$
              Now, notice that $$Au=(1-a)u+au (u^Tu)=(1-a)u+3au=(1+2a)u$$
              So $u$ is an eigenvector of $A$ with eigenvalue $(1+2a)$.



              Now, let's determine the other eigenvalues of $A$. Suppose $xinmathbb R^3$ is orthogonal to $u$, that is $u^Tx=0$, then
              $$Ax=(1-a)x+auu^Tx=(1-a)x$$
              so $x$ is an eigenvector of $A$ with eigenvalue $(1-a)$. Since the space of vectors $x$ orthogonal to $u$ is of dimension 2, we have now determined that $(1-a)$ is an eigenvalue of multiplicity 2 for $A$.



              In summary, the spectrum of $A$ is $(1+2a)$ (multiplicity 1) and $(1-a)$ (multiplicity 2). Consequently, $A$ is positive definite iff $$1+2a>0 text{ and } 1-a >0$$ that is iff $-1 < 2a < 2$$






              share|cite|improve this answer

























                up vote
                1
                down vote













                Let's determine the eigenvalues of $A$. If we can find the conditions for which those eigenvalues are all positive, then we can solve the question.



                Let $u in mathbb R^3$ be the vector with $1$ for all its components. Then note that $uu^T$ is the $3times 3$ matrix with $1$ for all its entries.
                Therefore $$A=(1-a)I+auu^T$$
                Now, notice that $$Au=(1-a)u+au (u^Tu)=(1-a)u+3au=(1+2a)u$$
                So $u$ is an eigenvector of $A$ with eigenvalue $(1+2a)$.



                Now, let's determine the other eigenvalues of $A$. Suppose $xinmathbb R^3$ is orthogonal to $u$, that is $u^Tx=0$, then
                $$Ax=(1-a)x+auu^Tx=(1-a)x$$
                so $x$ is an eigenvector of $A$ with eigenvalue $(1-a)$. Since the space of vectors $x$ orthogonal to $u$ is of dimension 2, we have now determined that $(1-a)$ is an eigenvalue of multiplicity 2 for $A$.



                In summary, the spectrum of $A$ is $(1+2a)$ (multiplicity 1) and $(1-a)$ (multiplicity 2). Consequently, $A$ is positive definite iff $$1+2a>0 text{ and } 1-a >0$$ that is iff $-1 < 2a < 2$$






                share|cite|improve this answer























                  up vote
                  1
                  down vote










                  up vote
                  1
                  down vote









                  Let's determine the eigenvalues of $A$. If we can find the conditions for which those eigenvalues are all positive, then we can solve the question.



                  Let $u in mathbb R^3$ be the vector with $1$ for all its components. Then note that $uu^T$ is the $3times 3$ matrix with $1$ for all its entries.
                  Therefore $$A=(1-a)I+auu^T$$
                  Now, notice that $$Au=(1-a)u+au (u^Tu)=(1-a)u+3au=(1+2a)u$$
                  So $u$ is an eigenvector of $A$ with eigenvalue $(1+2a)$.



                  Now, let's determine the other eigenvalues of $A$. Suppose $xinmathbb R^3$ is orthogonal to $u$, that is $u^Tx=0$, then
                  $$Ax=(1-a)x+auu^Tx=(1-a)x$$
                  so $x$ is an eigenvector of $A$ with eigenvalue $(1-a)$. Since the space of vectors $x$ orthogonal to $u$ is of dimension 2, we have now determined that $(1-a)$ is an eigenvalue of multiplicity 2 for $A$.



                  In summary, the spectrum of $A$ is $(1+2a)$ (multiplicity 1) and $(1-a)$ (multiplicity 2). Consequently, $A$ is positive definite iff $$1+2a>0 text{ and } 1-a >0$$ that is iff $-1 < 2a < 2$$






                  share|cite|improve this answer












                  Let's determine the eigenvalues of $A$. If we can find the conditions for which those eigenvalues are all positive, then we can solve the question.



                  Let $u in mathbb R^3$ be the vector with $1$ for all its components. Then note that $uu^T$ is the $3times 3$ matrix with $1$ for all its entries.
                  Therefore $$A=(1-a)I+auu^T$$
                  Now, notice that $$Au=(1-a)u+au (u^Tu)=(1-a)u+3au=(1+2a)u$$
                  So $u$ is an eigenvector of $A$ with eigenvalue $(1+2a)$.



                  Now, let's determine the other eigenvalues of $A$. Suppose $xinmathbb R^3$ is orthogonal to $u$, that is $u^Tx=0$, then
                  $$Ax=(1-a)x+auu^Tx=(1-a)x$$
                  so $x$ is an eigenvector of $A$ with eigenvalue $(1-a)$. Since the space of vectors $x$ orthogonal to $u$ is of dimension 2, we have now determined that $(1-a)$ is an eigenvalue of multiplicity 2 for $A$.



                  In summary, the spectrum of $A$ is $(1+2a)$ (multiplicity 1) and $(1-a)$ (multiplicity 2). Consequently, $A$ is positive definite iff $$1+2a>0 text{ and } 1-a >0$$ that is iff $-1 < 2a < 2$$







                  share|cite|improve this answer












                  share|cite|improve this answer



                  share|cite|improve this answer










                  answered Nov 13 at 18:24









                  Stefan Lafon

                  65716




                  65716






























                       

                      draft saved


                      draft discarded



















































                       


                      draft saved


                      draft discarded














                      StackExchange.ready(
                      function () {
                      StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2996981%2fconditions-for-a-certain-parametric-matrix-to-be-positive-definite%23new-answer', 'question_page');
                      }
                      );

                      Post as a guest















                      Required, but never shown





















































                      Required, but never shown














                      Required, but never shown












                      Required, but never shown







                      Required, but never shown

































                      Required, but never shown














                      Required, but never shown












                      Required, but never shown







                      Required, but never shown







                      Popular posts from this blog

                      How do I know what Microsoft account the skydrive app is syncing to?

                      Grease: Live!

                      When does type information flow backwards in C++?