How do I show this $xyleq frac{1}{p}x^p+frac{1}{q}y^q.$ without using concavity of log function?












3












$begingroup$


let reals $p, q >1$ and $x, y$ are positive real numbers with $1/p+1/q=1$ I want to show $xyleq frac{1}{p}x^p+frac{1}{q}y^q.$ without using concavity of $log$ as shown below



Proof with concavity of log function and exponential Growth



$lnleft(frac{1}{p}x^p+frac{1}{q}y^qright)geq frac{1}{p}lnleft(x^pright)+frac{1}{q}lnleft(y^qright)=lnleft(xyright).$



just to raise expon for both side we get the result .



then any other method ?










share|cite|improve this question









$endgroup$








  • 1




    $begingroup$
    Check proof 2 here: proofwiki.org/wiki/Young%27s_Inequality_for_Products
    $endgroup$
    – John11
    Dec 26 '18 at 22:49










  • $begingroup$
    Just to continue John11's comments, like this one
    $endgroup$
    – rtybase
    Dec 26 '18 at 23:56










  • $begingroup$
    Can you go between $(1-t)x+ty$ and $x/p+y/q$? It's pretty trivial, but usually needs to be pointed out in some minor capacity to be thought of.
    $endgroup$
    – Robert Wolfe
    Dec 27 '18 at 0:42
















3












$begingroup$


let reals $p, q >1$ and $x, y$ are positive real numbers with $1/p+1/q=1$ I want to show $xyleq frac{1}{p}x^p+frac{1}{q}y^q.$ without using concavity of $log$ as shown below



Proof with concavity of log function and exponential Growth



$lnleft(frac{1}{p}x^p+frac{1}{q}y^qright)geq frac{1}{p}lnleft(x^pright)+frac{1}{q}lnleft(y^qright)=lnleft(xyright).$



just to raise expon for both side we get the result .



then any other method ?










share|cite|improve this question









$endgroup$








  • 1




    $begingroup$
    Check proof 2 here: proofwiki.org/wiki/Young%27s_Inequality_for_Products
    $endgroup$
    – John11
    Dec 26 '18 at 22:49










  • $begingroup$
    Just to continue John11's comments, like this one
    $endgroup$
    – rtybase
    Dec 26 '18 at 23:56










  • $begingroup$
    Can you go between $(1-t)x+ty$ and $x/p+y/q$? It's pretty trivial, but usually needs to be pointed out in some minor capacity to be thought of.
    $endgroup$
    – Robert Wolfe
    Dec 27 '18 at 0:42














3












3








3


2



$begingroup$


let reals $p, q >1$ and $x, y$ are positive real numbers with $1/p+1/q=1$ I want to show $xyleq frac{1}{p}x^p+frac{1}{q}y^q.$ without using concavity of $log$ as shown below



Proof with concavity of log function and exponential Growth



$lnleft(frac{1}{p}x^p+frac{1}{q}y^qright)geq frac{1}{p}lnleft(x^pright)+frac{1}{q}lnleft(y^qright)=lnleft(xyright).$



just to raise expon for both side we get the result .



then any other method ?










share|cite|improve this question









$endgroup$




let reals $p, q >1$ and $x, y$ are positive real numbers with $1/p+1/q=1$ I want to show $xyleq frac{1}{p}x^p+frac{1}{q}y^q.$ without using concavity of $log$ as shown below



Proof with concavity of log function and exponential Growth



$lnleft(frac{1}{p}x^p+frac{1}{q}y^qright)geq frac{1}{p}lnleft(x^pright)+frac{1}{q}lnleft(y^qright)=lnleft(xyright).$



just to raise expon for both side we get the result .



then any other method ?







inequality convex-analysis holder-inequality






share|cite|improve this question













share|cite|improve this question











share|cite|improve this question




share|cite|improve this question










asked Dec 26 '18 at 22:42









zeraoulia rafikzeraoulia rafik

2,39711134




2,39711134








  • 1




    $begingroup$
    Check proof 2 here: proofwiki.org/wiki/Young%27s_Inequality_for_Products
    $endgroup$
    – John11
    Dec 26 '18 at 22:49










  • $begingroup$
    Just to continue John11's comments, like this one
    $endgroup$
    – rtybase
    Dec 26 '18 at 23:56










  • $begingroup$
    Can you go between $(1-t)x+ty$ and $x/p+y/q$? It's pretty trivial, but usually needs to be pointed out in some minor capacity to be thought of.
    $endgroup$
    – Robert Wolfe
    Dec 27 '18 at 0:42














  • 1




    $begingroup$
    Check proof 2 here: proofwiki.org/wiki/Young%27s_Inequality_for_Products
    $endgroup$
    – John11
    Dec 26 '18 at 22:49










  • $begingroup$
    Just to continue John11's comments, like this one
    $endgroup$
    – rtybase
    Dec 26 '18 at 23:56










  • $begingroup$
    Can you go between $(1-t)x+ty$ and $x/p+y/q$? It's pretty trivial, but usually needs to be pointed out in some minor capacity to be thought of.
    $endgroup$
    – Robert Wolfe
    Dec 27 '18 at 0:42








1




1




$begingroup$
Check proof 2 here: proofwiki.org/wiki/Young%27s_Inequality_for_Products
$endgroup$
– John11
Dec 26 '18 at 22:49




$begingroup$
Check proof 2 here: proofwiki.org/wiki/Young%27s_Inequality_for_Products
$endgroup$
– John11
Dec 26 '18 at 22:49












$begingroup$
Just to continue John11's comments, like this one
$endgroup$
– rtybase
Dec 26 '18 at 23:56




$begingroup$
Just to continue John11's comments, like this one
$endgroup$
– rtybase
Dec 26 '18 at 23:56












$begingroup$
Can you go between $(1-t)x+ty$ and $x/p+y/q$? It's pretty trivial, but usually needs to be pointed out in some minor capacity to be thought of.
$endgroup$
– Robert Wolfe
Dec 27 '18 at 0:42




$begingroup$
Can you go between $(1-t)x+ty$ and $x/p+y/q$? It's pretty trivial, but usually needs to be pointed out in some minor capacity to be thought of.
$endgroup$
– Robert Wolfe
Dec 27 '18 at 0:42










4 Answers
4






active

oldest

votes


















1












$begingroup$

Method 1:



For non-negative $u,v$ let $f(u)=u^{p-1}$ and $g(v)=v^{q-1}.$ Since $(p-1)(q-1)=1,$ the graphs of $(u,f(u))$ and of $(g(v),v)$ are the same graph, which I will call $G.$



Now $x^p/p=int_0^xf(u)du$ is the area of the figure bounded by $G$ and ${x}times [0,f(x)]$ and the $u$-axis in the $(u,v)$-plane. And $y^q/q=int_0^yg(y)dy$ is the area of the figure bounded by $G$ and $[0, g(y)]times {y}$ and the $v$-axis.



If $x^p=y^q,$ the union of these two figures is the rectangle $R(x,y)$ with corners $(0,0),(0,y),(x,y),(x,0),$ which has area $xy,$ so $x^p/p+y^q/q=int_0^xf(u)du+int_0^yg(v)dv=xy.$



If $x>y^{q/p}$ then the union of these two figures is $R(x,y)$ plus a region bounded by $G$, the $u$-axis, and the segments ${y^{q/p}}times [0,y]$ and ${x}times [0,f(x)],$ so $x^p/p+y^q/q=xy+int_{y^{q/p}}^x f(u)du>xy$.



Similarly if $x<y^{p/q}$ then $y>x^{q/p}$ and $x^p/p+y^q/q=xy+int_{x^{q/p}}^y g(v)dv>xy.$



This is easy to see on a diagram.



Generally, if $f:[0,infty)to [0,infty)$ is continuous and strictly monotonic, and $f(0)=0,$ then $xyleq int_0^xf(u)du+int_0^yf^{-1}(v)dv.$



Method 2:



For non-negative $x,y$ let $H(x,y)=x^p/p+y^q/q-xy.$ Then $frac {partial H(x,y)}{partial x}=frac1px^{p-1}-y,$ which is negative if $0leq x^{p-1}<y$ and positive if $x^{p-1}>y.$ So for a fixed $y,$ the function $H(x,y)$ has its minimum when $x^{p-1}=y.$ And we have $H(x, x^{p-1})=0.$






share|cite|improve this answer











$endgroup$





















    3












    $begingroup$

    This proof is from Steele's Cauchy-Schwarz Master Class. The weighted AM-GM inequality gives us
    begin{align*}
    u^alpha v^beta le frac{alpha}{alpha + beta}u^{alpha + beta} + frac{beta}{alpha + beta}v^{alpha + beta}
    end{align*}

    for $u, v ge 0$ and $alpha, beta > 0$. Substituting $x = u^alpha$, $y= v^beta$, $p = (alpha+beta)/alpha$, and $q = (alpha + beta)/beta$ gives us the desired inequality.






    share|cite|improve this answer









    $endgroup$





















      1












      $begingroup$

      Another cool proof is using Legendre's transform, with $L(x) = frac{1}{p}|x|^p$ is convex, then
      $$ L^*(y) = sup_{xin mathbb{R}} Big(xcdot y - L(y)Big) = frac{1}{q}|y|^q.$$
      Therefore
      $$ L(x)+L^*(y) = frac{1}{p}|x|^p + frac{1}{q}|y|^q geq xcdot{y}.$$



      More on this, see https://en.wikipedia.org/wiki/Legendre_transformation. It is also called "Fenchel's inequality".






      share|cite|improve this answer











      $endgroup$





















        0












        $begingroup$

        Hint



        Let $$f(t)=frac{t^p}{p}+frac{1}{q}-t text{ on } [0,infty)$$
        Now show that $t=1$ is a critical point and $f$ is an increasing function for $t geq 1$. This will imply
        $$f(t) geq f(1)=0.$$
        Consequently,
        $$frac{t^p}{p}+frac{1}{q} geq t$$
        Now let $t=xy^{frac{1}{p-1}}$.






        share|cite|improve this answer











        $endgroup$














          Your Answer








          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',
          autoActivateHeartbeat: false,
          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%2f3053388%2fhow-do-i-show-this-xy-leq-frac1pxp-frac1qyq-without-using-concavi%23new-answer', 'question_page');
          }
          );

          Post as a guest















          Required, but never shown

























          4 Answers
          4






          active

          oldest

          votes








          4 Answers
          4






          active

          oldest

          votes









          active

          oldest

          votes






          active

          oldest

          votes









          1












          $begingroup$

          Method 1:



          For non-negative $u,v$ let $f(u)=u^{p-1}$ and $g(v)=v^{q-1}.$ Since $(p-1)(q-1)=1,$ the graphs of $(u,f(u))$ and of $(g(v),v)$ are the same graph, which I will call $G.$



          Now $x^p/p=int_0^xf(u)du$ is the area of the figure bounded by $G$ and ${x}times [0,f(x)]$ and the $u$-axis in the $(u,v)$-plane. And $y^q/q=int_0^yg(y)dy$ is the area of the figure bounded by $G$ and $[0, g(y)]times {y}$ and the $v$-axis.



          If $x^p=y^q,$ the union of these two figures is the rectangle $R(x,y)$ with corners $(0,0),(0,y),(x,y),(x,0),$ which has area $xy,$ so $x^p/p+y^q/q=int_0^xf(u)du+int_0^yg(v)dv=xy.$



          If $x>y^{q/p}$ then the union of these two figures is $R(x,y)$ plus a region bounded by $G$, the $u$-axis, and the segments ${y^{q/p}}times [0,y]$ and ${x}times [0,f(x)],$ so $x^p/p+y^q/q=xy+int_{y^{q/p}}^x f(u)du>xy$.



          Similarly if $x<y^{p/q}$ then $y>x^{q/p}$ and $x^p/p+y^q/q=xy+int_{x^{q/p}}^y g(v)dv>xy.$



          This is easy to see on a diagram.



          Generally, if $f:[0,infty)to [0,infty)$ is continuous and strictly monotonic, and $f(0)=0,$ then $xyleq int_0^xf(u)du+int_0^yf^{-1}(v)dv.$



          Method 2:



          For non-negative $x,y$ let $H(x,y)=x^p/p+y^q/q-xy.$ Then $frac {partial H(x,y)}{partial x}=frac1px^{p-1}-y,$ which is negative if $0leq x^{p-1}<y$ and positive if $x^{p-1}>y.$ So for a fixed $y,$ the function $H(x,y)$ has its minimum when $x^{p-1}=y.$ And we have $H(x, x^{p-1})=0.$






          share|cite|improve this answer











          $endgroup$


















            1












            $begingroup$

            Method 1:



            For non-negative $u,v$ let $f(u)=u^{p-1}$ and $g(v)=v^{q-1}.$ Since $(p-1)(q-1)=1,$ the graphs of $(u,f(u))$ and of $(g(v),v)$ are the same graph, which I will call $G.$



            Now $x^p/p=int_0^xf(u)du$ is the area of the figure bounded by $G$ and ${x}times [0,f(x)]$ and the $u$-axis in the $(u,v)$-plane. And $y^q/q=int_0^yg(y)dy$ is the area of the figure bounded by $G$ and $[0, g(y)]times {y}$ and the $v$-axis.



            If $x^p=y^q,$ the union of these two figures is the rectangle $R(x,y)$ with corners $(0,0),(0,y),(x,y),(x,0),$ which has area $xy,$ so $x^p/p+y^q/q=int_0^xf(u)du+int_0^yg(v)dv=xy.$



            If $x>y^{q/p}$ then the union of these two figures is $R(x,y)$ plus a region bounded by $G$, the $u$-axis, and the segments ${y^{q/p}}times [0,y]$ and ${x}times [0,f(x)],$ so $x^p/p+y^q/q=xy+int_{y^{q/p}}^x f(u)du>xy$.



            Similarly if $x<y^{p/q}$ then $y>x^{q/p}$ and $x^p/p+y^q/q=xy+int_{x^{q/p}}^y g(v)dv>xy.$



            This is easy to see on a diagram.



            Generally, if $f:[0,infty)to [0,infty)$ is continuous and strictly monotonic, and $f(0)=0,$ then $xyleq int_0^xf(u)du+int_0^yf^{-1}(v)dv.$



            Method 2:



            For non-negative $x,y$ let $H(x,y)=x^p/p+y^q/q-xy.$ Then $frac {partial H(x,y)}{partial x}=frac1px^{p-1}-y,$ which is negative if $0leq x^{p-1}<y$ and positive if $x^{p-1}>y.$ So for a fixed $y,$ the function $H(x,y)$ has its minimum when $x^{p-1}=y.$ And we have $H(x, x^{p-1})=0.$






            share|cite|improve this answer











            $endgroup$
















              1












              1








              1





              $begingroup$

              Method 1:



              For non-negative $u,v$ let $f(u)=u^{p-1}$ and $g(v)=v^{q-1}.$ Since $(p-1)(q-1)=1,$ the graphs of $(u,f(u))$ and of $(g(v),v)$ are the same graph, which I will call $G.$



              Now $x^p/p=int_0^xf(u)du$ is the area of the figure bounded by $G$ and ${x}times [0,f(x)]$ and the $u$-axis in the $(u,v)$-plane. And $y^q/q=int_0^yg(y)dy$ is the area of the figure bounded by $G$ and $[0, g(y)]times {y}$ and the $v$-axis.



              If $x^p=y^q,$ the union of these two figures is the rectangle $R(x,y)$ with corners $(0,0),(0,y),(x,y),(x,0),$ which has area $xy,$ so $x^p/p+y^q/q=int_0^xf(u)du+int_0^yg(v)dv=xy.$



              If $x>y^{q/p}$ then the union of these two figures is $R(x,y)$ plus a region bounded by $G$, the $u$-axis, and the segments ${y^{q/p}}times [0,y]$ and ${x}times [0,f(x)],$ so $x^p/p+y^q/q=xy+int_{y^{q/p}}^x f(u)du>xy$.



              Similarly if $x<y^{p/q}$ then $y>x^{q/p}$ and $x^p/p+y^q/q=xy+int_{x^{q/p}}^y g(v)dv>xy.$



              This is easy to see on a diagram.



              Generally, if $f:[0,infty)to [0,infty)$ is continuous and strictly monotonic, and $f(0)=0,$ then $xyleq int_0^xf(u)du+int_0^yf^{-1}(v)dv.$



              Method 2:



              For non-negative $x,y$ let $H(x,y)=x^p/p+y^q/q-xy.$ Then $frac {partial H(x,y)}{partial x}=frac1px^{p-1}-y,$ which is negative if $0leq x^{p-1}<y$ and positive if $x^{p-1}>y.$ So for a fixed $y,$ the function $H(x,y)$ has its minimum when $x^{p-1}=y.$ And we have $H(x, x^{p-1})=0.$






              share|cite|improve this answer











              $endgroup$



              Method 1:



              For non-negative $u,v$ let $f(u)=u^{p-1}$ and $g(v)=v^{q-1}.$ Since $(p-1)(q-1)=1,$ the graphs of $(u,f(u))$ and of $(g(v),v)$ are the same graph, which I will call $G.$



              Now $x^p/p=int_0^xf(u)du$ is the area of the figure bounded by $G$ and ${x}times [0,f(x)]$ and the $u$-axis in the $(u,v)$-plane. And $y^q/q=int_0^yg(y)dy$ is the area of the figure bounded by $G$ and $[0, g(y)]times {y}$ and the $v$-axis.



              If $x^p=y^q,$ the union of these two figures is the rectangle $R(x,y)$ with corners $(0,0),(0,y),(x,y),(x,0),$ which has area $xy,$ so $x^p/p+y^q/q=int_0^xf(u)du+int_0^yg(v)dv=xy.$



              If $x>y^{q/p}$ then the union of these two figures is $R(x,y)$ plus a region bounded by $G$, the $u$-axis, and the segments ${y^{q/p}}times [0,y]$ and ${x}times [0,f(x)],$ so $x^p/p+y^q/q=xy+int_{y^{q/p}}^x f(u)du>xy$.



              Similarly if $x<y^{p/q}$ then $y>x^{q/p}$ and $x^p/p+y^q/q=xy+int_{x^{q/p}}^y g(v)dv>xy.$



              This is easy to see on a diagram.



              Generally, if $f:[0,infty)to [0,infty)$ is continuous and strictly monotonic, and $f(0)=0,$ then $xyleq int_0^xf(u)du+int_0^yf^{-1}(v)dv.$



              Method 2:



              For non-negative $x,y$ let $H(x,y)=x^p/p+y^q/q-xy.$ Then $frac {partial H(x,y)}{partial x}=frac1px^{p-1}-y,$ which is negative if $0leq x^{p-1}<y$ and positive if $x^{p-1}>y.$ So for a fixed $y,$ the function $H(x,y)$ has its minimum when $x^{p-1}=y.$ And we have $H(x, x^{p-1})=0.$







              share|cite|improve this answer














              share|cite|improve this answer



              share|cite|improve this answer








              edited Dec 27 '18 at 10:25









              zeraoulia rafik

              2,39711134




              2,39711134










              answered Dec 27 '18 at 9:51









              DanielWainfleetDanielWainfleet

              35.8k31648




              35.8k31648























                  3












                  $begingroup$

                  This proof is from Steele's Cauchy-Schwarz Master Class. The weighted AM-GM inequality gives us
                  begin{align*}
                  u^alpha v^beta le frac{alpha}{alpha + beta}u^{alpha + beta} + frac{beta}{alpha + beta}v^{alpha + beta}
                  end{align*}

                  for $u, v ge 0$ and $alpha, beta > 0$. Substituting $x = u^alpha$, $y= v^beta$, $p = (alpha+beta)/alpha$, and $q = (alpha + beta)/beta$ gives us the desired inequality.






                  share|cite|improve this answer









                  $endgroup$


















                    3












                    $begingroup$

                    This proof is from Steele's Cauchy-Schwarz Master Class. The weighted AM-GM inequality gives us
                    begin{align*}
                    u^alpha v^beta le frac{alpha}{alpha + beta}u^{alpha + beta} + frac{beta}{alpha + beta}v^{alpha + beta}
                    end{align*}

                    for $u, v ge 0$ and $alpha, beta > 0$. Substituting $x = u^alpha$, $y= v^beta$, $p = (alpha+beta)/alpha$, and $q = (alpha + beta)/beta$ gives us the desired inequality.






                    share|cite|improve this answer









                    $endgroup$
















                      3












                      3








                      3





                      $begingroup$

                      This proof is from Steele's Cauchy-Schwarz Master Class. The weighted AM-GM inequality gives us
                      begin{align*}
                      u^alpha v^beta le frac{alpha}{alpha + beta}u^{alpha + beta} + frac{beta}{alpha + beta}v^{alpha + beta}
                      end{align*}

                      for $u, v ge 0$ and $alpha, beta > 0$. Substituting $x = u^alpha$, $y= v^beta$, $p = (alpha+beta)/alpha$, and $q = (alpha + beta)/beta$ gives us the desired inequality.






                      share|cite|improve this answer









                      $endgroup$



                      This proof is from Steele's Cauchy-Schwarz Master Class. The weighted AM-GM inequality gives us
                      begin{align*}
                      u^alpha v^beta le frac{alpha}{alpha + beta}u^{alpha + beta} + frac{beta}{alpha + beta}v^{alpha + beta}
                      end{align*}

                      for $u, v ge 0$ and $alpha, beta > 0$. Substituting $x = u^alpha$, $y= v^beta$, $p = (alpha+beta)/alpha$, and $q = (alpha + beta)/beta$ gives us the desired inequality.







                      share|cite|improve this answer












                      share|cite|improve this answer



                      share|cite|improve this answer










                      answered Dec 26 '18 at 22:47









                      Tom ChenTom Chen

                      2,068715




                      2,068715























                          1












                          $begingroup$

                          Another cool proof is using Legendre's transform, with $L(x) = frac{1}{p}|x|^p$ is convex, then
                          $$ L^*(y) = sup_{xin mathbb{R}} Big(xcdot y - L(y)Big) = frac{1}{q}|y|^q.$$
                          Therefore
                          $$ L(x)+L^*(y) = frac{1}{p}|x|^p + frac{1}{q}|y|^q geq xcdot{y}.$$



                          More on this, see https://en.wikipedia.org/wiki/Legendre_transformation. It is also called "Fenchel's inequality".






                          share|cite|improve this answer











                          $endgroup$


















                            1












                            $begingroup$

                            Another cool proof is using Legendre's transform, with $L(x) = frac{1}{p}|x|^p$ is convex, then
                            $$ L^*(y) = sup_{xin mathbb{R}} Big(xcdot y - L(y)Big) = frac{1}{q}|y|^q.$$
                            Therefore
                            $$ L(x)+L^*(y) = frac{1}{p}|x|^p + frac{1}{q}|y|^q geq xcdot{y}.$$



                            More on this, see https://en.wikipedia.org/wiki/Legendre_transformation. It is also called "Fenchel's inequality".






                            share|cite|improve this answer











                            $endgroup$
















                              1












                              1








                              1





                              $begingroup$

                              Another cool proof is using Legendre's transform, with $L(x) = frac{1}{p}|x|^p$ is convex, then
                              $$ L^*(y) = sup_{xin mathbb{R}} Big(xcdot y - L(y)Big) = frac{1}{q}|y|^q.$$
                              Therefore
                              $$ L(x)+L^*(y) = frac{1}{p}|x|^p + frac{1}{q}|y|^q geq xcdot{y}.$$



                              More on this, see https://en.wikipedia.org/wiki/Legendre_transformation. It is also called "Fenchel's inequality".






                              share|cite|improve this answer











                              $endgroup$



                              Another cool proof is using Legendre's transform, with $L(x) = frac{1}{p}|x|^p$ is convex, then
                              $$ L^*(y) = sup_{xin mathbb{R}} Big(xcdot y - L(y)Big) = frac{1}{q}|y|^q.$$
                              Therefore
                              $$ L(x)+L^*(y) = frac{1}{p}|x|^p + frac{1}{q}|y|^q geq xcdot{y}.$$



                              More on this, see https://en.wikipedia.org/wiki/Legendre_transformation. It is also called "Fenchel's inequality".







                              share|cite|improve this answer














                              share|cite|improve this answer



                              share|cite|improve this answer








                              edited Dec 27 '18 at 7:37

























                              answered Dec 27 '18 at 7:31









                              SeanSean

                              532513




                              532513























                                  0












                                  $begingroup$

                                  Hint



                                  Let $$f(t)=frac{t^p}{p}+frac{1}{q}-t text{ on } [0,infty)$$
                                  Now show that $t=1$ is a critical point and $f$ is an increasing function for $t geq 1$. This will imply
                                  $$f(t) geq f(1)=0.$$
                                  Consequently,
                                  $$frac{t^p}{p}+frac{1}{q} geq t$$
                                  Now let $t=xy^{frac{1}{p-1}}$.






                                  share|cite|improve this answer











                                  $endgroup$


















                                    0












                                    $begingroup$

                                    Hint



                                    Let $$f(t)=frac{t^p}{p}+frac{1}{q}-t text{ on } [0,infty)$$
                                    Now show that $t=1$ is a critical point and $f$ is an increasing function for $t geq 1$. This will imply
                                    $$f(t) geq f(1)=0.$$
                                    Consequently,
                                    $$frac{t^p}{p}+frac{1}{q} geq t$$
                                    Now let $t=xy^{frac{1}{p-1}}$.






                                    share|cite|improve this answer











                                    $endgroup$
















                                      0












                                      0








                                      0





                                      $begingroup$

                                      Hint



                                      Let $$f(t)=frac{t^p}{p}+frac{1}{q}-t text{ on } [0,infty)$$
                                      Now show that $t=1$ is a critical point and $f$ is an increasing function for $t geq 1$. This will imply
                                      $$f(t) geq f(1)=0.$$
                                      Consequently,
                                      $$frac{t^p}{p}+frac{1}{q} geq t$$
                                      Now let $t=xy^{frac{1}{p-1}}$.






                                      share|cite|improve this answer











                                      $endgroup$



                                      Hint



                                      Let $$f(t)=frac{t^p}{p}+frac{1}{q}-t text{ on } [0,infty)$$
                                      Now show that $t=1$ is a critical point and $f$ is an increasing function for $t geq 1$. This will imply
                                      $$f(t) geq f(1)=0.$$
                                      Consequently,
                                      $$frac{t^p}{p}+frac{1}{q} geq t$$
                                      Now let $t=xy^{frac{1}{p-1}}$.







                                      share|cite|improve this answer














                                      share|cite|improve this answer



                                      share|cite|improve this answer








                                      edited Dec 27 '18 at 20:45

























                                      answered Dec 26 '18 at 22:46









                                      Anurag AAnurag A

                                      26.4k12351




                                      26.4k12351






























                                          draft saved

                                          draft discarded




















































                                          Thanks for contributing an answer to Mathematics Stack Exchange!


                                          • Please be sure to answer the question. Provide details and share your research!

                                          But avoid



                                          • Asking for help, clarification, or responding to other answers.

                                          • Making statements based on opinion; back them up with references or personal experience.


                                          Use MathJax to format equations. MathJax reference.


                                          To learn more, see our tips on writing great answers.




                                          draft saved


                                          draft discarded














                                          StackExchange.ready(
                                          function () {
                                          StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3053388%2fhow-do-i-show-this-xy-leq-frac1pxp-frac1qyq-without-using-concavi%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

                                          Le Mesnil-Réaume

                                          Ida-Boy-Ed-Garten

                                          web3.py web3.isConnected() returns false always