Prove that $({aover a+b})^3+({bover b+c})^3+ ({cover c+a})^3geq {3over 8}$












11












$begingroup$



Let $a,b,c$ be positive real numbers. Prove that $$Big({aover a+b}Big)^3+Big({bover b+c}Big)^3+ Big({cover c+a}Big)^3geq {3over 8}$$




If we put $x=b/a$, $y= c/b$ and $z=a/c$ we get $xyz=1$ and



$$Big({1over 1+x}Big)^3+Big({1over 1+y}Big)^3+ Big({1over 1+z}Big)^3geq {3over 8}$$



Since $f(x)=Big({1over 1+x}Big)^3$ is convex we get, by Jensen,: $$Big({1over 1+x}Big)^3+Big({1over 1+y}Big)^3+ Big({1over 1+z}Big)^3geq 3f({x+y+zover 3})$$



Unfortunately, since $f$ is decreasing we don't have $f({x+y+zover 3}) geq f(1) = {1over 8}$.



Some idea how to solve this?










share|cite|improve this question











$endgroup$








  • 4




    $begingroup$
    How is this question "seeking personal advice," whoever voted to close this thread?
    $endgroup$
    – Batominovski
    Dec 4 '18 at 21:26


















11












$begingroup$



Let $a,b,c$ be positive real numbers. Prove that $$Big({aover a+b}Big)^3+Big({bover b+c}Big)^3+ Big({cover c+a}Big)^3geq {3over 8}$$




If we put $x=b/a$, $y= c/b$ and $z=a/c$ we get $xyz=1$ and



$$Big({1over 1+x}Big)^3+Big({1over 1+y}Big)^3+ Big({1over 1+z}Big)^3geq {3over 8}$$



Since $f(x)=Big({1over 1+x}Big)^3$ is convex we get, by Jensen,: $$Big({1over 1+x}Big)^3+Big({1over 1+y}Big)^3+ Big({1over 1+z}Big)^3geq 3f({x+y+zover 3})$$



Unfortunately, since $f$ is decreasing we don't have $f({x+y+zover 3}) geq f(1) = {1over 8}$.



Some idea how to solve this?










share|cite|improve this question











$endgroup$








  • 4




    $begingroup$
    How is this question "seeking personal advice," whoever voted to close this thread?
    $endgroup$
    – Batominovski
    Dec 4 '18 at 21:26
















11












11








11


4



$begingroup$



Let $a,b,c$ be positive real numbers. Prove that $$Big({aover a+b}Big)^3+Big({bover b+c}Big)^3+ Big({cover c+a}Big)^3geq {3over 8}$$




If we put $x=b/a$, $y= c/b$ and $z=a/c$ we get $xyz=1$ and



$$Big({1over 1+x}Big)^3+Big({1over 1+y}Big)^3+ Big({1over 1+z}Big)^3geq {3over 8}$$



Since $f(x)=Big({1over 1+x}Big)^3$ is convex we get, by Jensen,: $$Big({1over 1+x}Big)^3+Big({1over 1+y}Big)^3+ Big({1over 1+z}Big)^3geq 3f({x+y+zover 3})$$



Unfortunately, since $f$ is decreasing we don't have $f({x+y+zover 3}) geq f(1) = {1over 8}$.



Some idea how to solve this?










share|cite|improve this question











$endgroup$





Let $a,b,c$ be positive real numbers. Prove that $$Big({aover a+b}Big)^3+Big({bover b+c}Big)^3+ Big({cover c+a}Big)^3geq {3over 8}$$




If we put $x=b/a$, $y= c/b$ and $z=a/c$ we get $xyz=1$ and



$$Big({1over 1+x}Big)^3+Big({1over 1+y}Big)^3+ Big({1over 1+z}Big)^3geq {3over 8}$$



Since $f(x)=Big({1over 1+x}Big)^3$ is convex we get, by Jensen,: $$Big({1over 1+x}Big)^3+Big({1over 1+y}Big)^3+ Big({1over 1+z}Big)^3geq 3f({x+y+zover 3})$$



Unfortunately, since $f$ is decreasing we don't have $f({x+y+zover 3}) geq f(1) = {1over 8}$.



Some idea how to solve this?







inequality cauchy-schwarz-inequality holder-inequality jensen-inequality uvw






share|cite|improve this question















share|cite|improve this question













share|cite|improve this question




share|cite|improve this question








edited Dec 11 '18 at 19:51









Xander Henderson

14.3k103554




14.3k103554










asked Dec 4 '18 at 16:51









greedoidgreedoid

40.7k1149100




40.7k1149100








  • 4




    $begingroup$
    How is this question "seeking personal advice," whoever voted to close this thread?
    $endgroup$
    – Batominovski
    Dec 4 '18 at 21:26
















  • 4




    $begingroup$
    How is this question "seeking personal advice," whoever voted to close this thread?
    $endgroup$
    – Batominovski
    Dec 4 '18 at 21:26










4




4




$begingroup$
How is this question "seeking personal advice," whoever voted to close this thread?
$endgroup$
– Batominovski
Dec 4 '18 at 21:26






$begingroup$
How is this question "seeking personal advice," whoever voted to close this thread?
$endgroup$
– Batominovski
Dec 4 '18 at 21:26












3 Answers
3






active

oldest

votes


















4












$begingroup$

Let $g(t):=left(dfrac{1}{1+exp(t)}right)^3$ for $tinmathbb{R}$. Then, $$g''(t)=frac{3,exp(t),big(3,exp(t)-1big)}{big(1+exp(t)big)^5}text{ for each }tinmathbb{R},.$$
Thus, $g$ is convex on $big[-ln(3),inftybig)$.



Let $x:=dfrac{b}{a}$, $y:=dfrac{c}{b}$, and $z:=dfrac{a}{c}$ be as the OP defines. Then, the required inequality is equivalent to
$$gbig(ln(x)big)+gbig(ln(y)big)+gbig(ln(z)big)geq dfrac{3}{8},.tag{*}$$
If $x$, $y$, or $z$ is less than $dfrac{1}{3}$, then clearly the left-hand side of (*) is greater than
$$left(dfrac{1}{1+frac13}right)^3=frac{27}{64}>frac38,.$$
If all $x$, $y$, and $z$ are greater than or equal to $dfrac13$, then $ln(x),ln(y),ln(z)geq -ln(3)$, so that we can use convexity of $g$ on $big[-ln(3),inftybig)$. By Jensen's Inequality,
$$gbig(ln(x)big)+gbig(ln(y)big)+gbig(ln(z)big)geq3,gleft(frac{ln(x)+ln(y)+ln(z)}{3}right)=3,g(0)=frac{3}{8},.$$
Hence, the equality holds if and only if $x=y=z=1$, making $a=b=c$.






share|cite|improve this answer











$endgroup$













  • $begingroup$
    I don't get it. What is the difference with my solution?
    $endgroup$
    – greedoid
    Dec 4 '18 at 21:45










  • $begingroup$
    I'm not sure what to say to that, but my idea was to bypass the inequality $frac{x+y+z}{3}geq sqrt[3]{xyz}$ that you would need to use in your attempt.
    $endgroup$
    – Batominovski
    Dec 4 '18 at 21:48





















5












$begingroup$

By Holder $$left(sum_{cyc}frac{a^3}{(a+b)^3}right)^2sum_{cyc}1geqleft(sum_{cyc}sqrt[3]{left(frac{a^3}{(a+b)^3}right)^2cdot1}right)^3=left(sum_{cyc}frac{a^2}{(a+b)^2}right)^3.$$
Thus, it's enough to prove that
$$frac{left(sumlimits_{cyc}frac{a^2}{(a+b)^2}right)^3}{3}geqfrac{9}{64}$$ or
$$sumlimits_{cyc}frac{a^2}{(a+b)^2}geqfrac{3}{4}.$$
Now, by C-S
$$sumlimits_{cyc}frac{a^2}{(a+b)^2}=sumlimits_{cyc}frac{a^2(a+c)^2}{(a+b)^2(a+c)^2}geqfrac{left(sumlimits_{cyc}(a^2+ab)right)^2}{sumlimits_{cyc}(a+b)^2(a+c)^2}.$$
Thus, it's enough to prove that
$$4left(sumlimits_{cyc}(a^2+ab)right)^2geq3sumlimits_{cyc}(a+b)^2(a+c)^2,$$
which is true even for all reals $a$, $b$ and $c$.



Indeed, the last inequality is symmetric inequality by degree four,



which says that by $uvw$ (https://math.stackexchange.com/tags/uvw/info )



it's enough to prove the last inequality for equality case of two variables and since



it's the homogeneous inequality by even degree, we can assume $b=c=1$, which gives
$$(a-1)^2(a+3)^2geq0.$$
Done!






share|cite|improve this answer









$endgroup$





















    5












    $begingroup$

    This is more of a comment, but I don't have the reputation. Use Lagrange multipliers. Solving, you find that the critical points occur when $xyz=1$ and $yz (1+x)^4 = xz (1+y)^4 = xy (1+z)^4$. I think the only solution is $x=y=z=1$. Clearly, it's a minimum and plugging in shows the bound.



    Added:



    We can rewrite the condition as $xyz=1$ and $frac{(1+x)^4}{x} = frac{(1+y)^4}{y} = frac{(1+z)^4}{z}$



    The function $g(x)=frac{(1+x)^4}{x}$ is decreasing from $0$ to $1/3$ and increasing from $1/3$ to $infty$. This shows that 2 of $x,y,z$ must be equal (WLOG $x$ and $y$) and $z=1/x^2$. It remains to solve $frac{(1+x)^4}{x} = frac{(1+1/x^2)^4}{1/x^2}$. This time, it's not hard to check $x=1$ is the only solution and we are done.






    share|cite|improve this answer











    $endgroup$









    • 2




      $begingroup$
      It seems to me there's enough here to post as an answer, no need to apologize.
      $endgroup$
      – David K
      Dec 4 '18 at 21:11











    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',
    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%2f3025819%2fprove-that-a-over-ab3b-over-bc3-c-over-ca3-geq-3-over-8%23new-answer', 'question_page');
    }
    );

    Post as a guest















    Required, but never shown

























    3 Answers
    3






    active

    oldest

    votes








    3 Answers
    3






    active

    oldest

    votes









    active

    oldest

    votes






    active

    oldest

    votes









    4












    $begingroup$

    Let $g(t):=left(dfrac{1}{1+exp(t)}right)^3$ for $tinmathbb{R}$. Then, $$g''(t)=frac{3,exp(t),big(3,exp(t)-1big)}{big(1+exp(t)big)^5}text{ for each }tinmathbb{R},.$$
    Thus, $g$ is convex on $big[-ln(3),inftybig)$.



    Let $x:=dfrac{b}{a}$, $y:=dfrac{c}{b}$, and $z:=dfrac{a}{c}$ be as the OP defines. Then, the required inequality is equivalent to
    $$gbig(ln(x)big)+gbig(ln(y)big)+gbig(ln(z)big)geq dfrac{3}{8},.tag{*}$$
    If $x$, $y$, or $z$ is less than $dfrac{1}{3}$, then clearly the left-hand side of (*) is greater than
    $$left(dfrac{1}{1+frac13}right)^3=frac{27}{64}>frac38,.$$
    If all $x$, $y$, and $z$ are greater than or equal to $dfrac13$, then $ln(x),ln(y),ln(z)geq -ln(3)$, so that we can use convexity of $g$ on $big[-ln(3),inftybig)$. By Jensen's Inequality,
    $$gbig(ln(x)big)+gbig(ln(y)big)+gbig(ln(z)big)geq3,gleft(frac{ln(x)+ln(y)+ln(z)}{3}right)=3,g(0)=frac{3}{8},.$$
    Hence, the equality holds if and only if $x=y=z=1$, making $a=b=c$.






    share|cite|improve this answer











    $endgroup$













    • $begingroup$
      I don't get it. What is the difference with my solution?
      $endgroup$
      – greedoid
      Dec 4 '18 at 21:45










    • $begingroup$
      I'm not sure what to say to that, but my idea was to bypass the inequality $frac{x+y+z}{3}geq sqrt[3]{xyz}$ that you would need to use in your attempt.
      $endgroup$
      – Batominovski
      Dec 4 '18 at 21:48


















    4












    $begingroup$

    Let $g(t):=left(dfrac{1}{1+exp(t)}right)^3$ for $tinmathbb{R}$. Then, $$g''(t)=frac{3,exp(t),big(3,exp(t)-1big)}{big(1+exp(t)big)^5}text{ for each }tinmathbb{R},.$$
    Thus, $g$ is convex on $big[-ln(3),inftybig)$.



    Let $x:=dfrac{b}{a}$, $y:=dfrac{c}{b}$, and $z:=dfrac{a}{c}$ be as the OP defines. Then, the required inequality is equivalent to
    $$gbig(ln(x)big)+gbig(ln(y)big)+gbig(ln(z)big)geq dfrac{3}{8},.tag{*}$$
    If $x$, $y$, or $z$ is less than $dfrac{1}{3}$, then clearly the left-hand side of (*) is greater than
    $$left(dfrac{1}{1+frac13}right)^3=frac{27}{64}>frac38,.$$
    If all $x$, $y$, and $z$ are greater than or equal to $dfrac13$, then $ln(x),ln(y),ln(z)geq -ln(3)$, so that we can use convexity of $g$ on $big[-ln(3),inftybig)$. By Jensen's Inequality,
    $$gbig(ln(x)big)+gbig(ln(y)big)+gbig(ln(z)big)geq3,gleft(frac{ln(x)+ln(y)+ln(z)}{3}right)=3,g(0)=frac{3}{8},.$$
    Hence, the equality holds if and only if $x=y=z=1$, making $a=b=c$.






    share|cite|improve this answer











    $endgroup$













    • $begingroup$
      I don't get it. What is the difference with my solution?
      $endgroup$
      – greedoid
      Dec 4 '18 at 21:45










    • $begingroup$
      I'm not sure what to say to that, but my idea was to bypass the inequality $frac{x+y+z}{3}geq sqrt[3]{xyz}$ that you would need to use in your attempt.
      $endgroup$
      – Batominovski
      Dec 4 '18 at 21:48
















    4












    4








    4





    $begingroup$

    Let $g(t):=left(dfrac{1}{1+exp(t)}right)^3$ for $tinmathbb{R}$. Then, $$g''(t)=frac{3,exp(t),big(3,exp(t)-1big)}{big(1+exp(t)big)^5}text{ for each }tinmathbb{R},.$$
    Thus, $g$ is convex on $big[-ln(3),inftybig)$.



    Let $x:=dfrac{b}{a}$, $y:=dfrac{c}{b}$, and $z:=dfrac{a}{c}$ be as the OP defines. Then, the required inequality is equivalent to
    $$gbig(ln(x)big)+gbig(ln(y)big)+gbig(ln(z)big)geq dfrac{3}{8},.tag{*}$$
    If $x$, $y$, or $z$ is less than $dfrac{1}{3}$, then clearly the left-hand side of (*) is greater than
    $$left(dfrac{1}{1+frac13}right)^3=frac{27}{64}>frac38,.$$
    If all $x$, $y$, and $z$ are greater than or equal to $dfrac13$, then $ln(x),ln(y),ln(z)geq -ln(3)$, so that we can use convexity of $g$ on $big[-ln(3),inftybig)$. By Jensen's Inequality,
    $$gbig(ln(x)big)+gbig(ln(y)big)+gbig(ln(z)big)geq3,gleft(frac{ln(x)+ln(y)+ln(z)}{3}right)=3,g(0)=frac{3}{8},.$$
    Hence, the equality holds if and only if $x=y=z=1$, making $a=b=c$.






    share|cite|improve this answer











    $endgroup$



    Let $g(t):=left(dfrac{1}{1+exp(t)}right)^3$ for $tinmathbb{R}$. Then, $$g''(t)=frac{3,exp(t),big(3,exp(t)-1big)}{big(1+exp(t)big)^5}text{ for each }tinmathbb{R},.$$
    Thus, $g$ is convex on $big[-ln(3),inftybig)$.



    Let $x:=dfrac{b}{a}$, $y:=dfrac{c}{b}$, and $z:=dfrac{a}{c}$ be as the OP defines. Then, the required inequality is equivalent to
    $$gbig(ln(x)big)+gbig(ln(y)big)+gbig(ln(z)big)geq dfrac{3}{8},.tag{*}$$
    If $x$, $y$, or $z$ is less than $dfrac{1}{3}$, then clearly the left-hand side of (*) is greater than
    $$left(dfrac{1}{1+frac13}right)^3=frac{27}{64}>frac38,.$$
    If all $x$, $y$, and $z$ are greater than or equal to $dfrac13$, then $ln(x),ln(y),ln(z)geq -ln(3)$, so that we can use convexity of $g$ on $big[-ln(3),inftybig)$. By Jensen's Inequality,
    $$gbig(ln(x)big)+gbig(ln(y)big)+gbig(ln(z)big)geq3,gleft(frac{ln(x)+ln(y)+ln(z)}{3}right)=3,g(0)=frac{3}{8},.$$
    Hence, the equality holds if and only if $x=y=z=1$, making $a=b=c$.







    share|cite|improve this answer














    share|cite|improve this answer



    share|cite|improve this answer








    edited Dec 4 '18 at 21:46

























    answered Dec 4 '18 at 21:42









    BatominovskiBatominovski

    1




    1












    • $begingroup$
      I don't get it. What is the difference with my solution?
      $endgroup$
      – greedoid
      Dec 4 '18 at 21:45










    • $begingroup$
      I'm not sure what to say to that, but my idea was to bypass the inequality $frac{x+y+z}{3}geq sqrt[3]{xyz}$ that you would need to use in your attempt.
      $endgroup$
      – Batominovski
      Dec 4 '18 at 21:48




















    • $begingroup$
      I don't get it. What is the difference with my solution?
      $endgroup$
      – greedoid
      Dec 4 '18 at 21:45










    • $begingroup$
      I'm not sure what to say to that, but my idea was to bypass the inequality $frac{x+y+z}{3}geq sqrt[3]{xyz}$ that you would need to use in your attempt.
      $endgroup$
      – Batominovski
      Dec 4 '18 at 21:48


















    $begingroup$
    I don't get it. What is the difference with my solution?
    $endgroup$
    – greedoid
    Dec 4 '18 at 21:45




    $begingroup$
    I don't get it. What is the difference with my solution?
    $endgroup$
    – greedoid
    Dec 4 '18 at 21:45












    $begingroup$
    I'm not sure what to say to that, but my idea was to bypass the inequality $frac{x+y+z}{3}geq sqrt[3]{xyz}$ that you would need to use in your attempt.
    $endgroup$
    – Batominovski
    Dec 4 '18 at 21:48






    $begingroup$
    I'm not sure what to say to that, but my idea was to bypass the inequality $frac{x+y+z}{3}geq sqrt[3]{xyz}$ that you would need to use in your attempt.
    $endgroup$
    – Batominovski
    Dec 4 '18 at 21:48













    5












    $begingroup$

    By Holder $$left(sum_{cyc}frac{a^3}{(a+b)^3}right)^2sum_{cyc}1geqleft(sum_{cyc}sqrt[3]{left(frac{a^3}{(a+b)^3}right)^2cdot1}right)^3=left(sum_{cyc}frac{a^2}{(a+b)^2}right)^3.$$
    Thus, it's enough to prove that
    $$frac{left(sumlimits_{cyc}frac{a^2}{(a+b)^2}right)^3}{3}geqfrac{9}{64}$$ or
    $$sumlimits_{cyc}frac{a^2}{(a+b)^2}geqfrac{3}{4}.$$
    Now, by C-S
    $$sumlimits_{cyc}frac{a^2}{(a+b)^2}=sumlimits_{cyc}frac{a^2(a+c)^2}{(a+b)^2(a+c)^2}geqfrac{left(sumlimits_{cyc}(a^2+ab)right)^2}{sumlimits_{cyc}(a+b)^2(a+c)^2}.$$
    Thus, it's enough to prove that
    $$4left(sumlimits_{cyc}(a^2+ab)right)^2geq3sumlimits_{cyc}(a+b)^2(a+c)^2,$$
    which is true even for all reals $a$, $b$ and $c$.



    Indeed, the last inequality is symmetric inequality by degree four,



    which says that by $uvw$ (https://math.stackexchange.com/tags/uvw/info )



    it's enough to prove the last inequality for equality case of two variables and since



    it's the homogeneous inequality by even degree, we can assume $b=c=1$, which gives
    $$(a-1)^2(a+3)^2geq0.$$
    Done!






    share|cite|improve this answer









    $endgroup$


















      5












      $begingroup$

      By Holder $$left(sum_{cyc}frac{a^3}{(a+b)^3}right)^2sum_{cyc}1geqleft(sum_{cyc}sqrt[3]{left(frac{a^3}{(a+b)^3}right)^2cdot1}right)^3=left(sum_{cyc}frac{a^2}{(a+b)^2}right)^3.$$
      Thus, it's enough to prove that
      $$frac{left(sumlimits_{cyc}frac{a^2}{(a+b)^2}right)^3}{3}geqfrac{9}{64}$$ or
      $$sumlimits_{cyc}frac{a^2}{(a+b)^2}geqfrac{3}{4}.$$
      Now, by C-S
      $$sumlimits_{cyc}frac{a^2}{(a+b)^2}=sumlimits_{cyc}frac{a^2(a+c)^2}{(a+b)^2(a+c)^2}geqfrac{left(sumlimits_{cyc}(a^2+ab)right)^2}{sumlimits_{cyc}(a+b)^2(a+c)^2}.$$
      Thus, it's enough to prove that
      $$4left(sumlimits_{cyc}(a^2+ab)right)^2geq3sumlimits_{cyc}(a+b)^2(a+c)^2,$$
      which is true even for all reals $a$, $b$ and $c$.



      Indeed, the last inequality is symmetric inequality by degree four,



      which says that by $uvw$ (https://math.stackexchange.com/tags/uvw/info )



      it's enough to prove the last inequality for equality case of two variables and since



      it's the homogeneous inequality by even degree, we can assume $b=c=1$, which gives
      $$(a-1)^2(a+3)^2geq0.$$
      Done!






      share|cite|improve this answer









      $endgroup$
















        5












        5








        5





        $begingroup$

        By Holder $$left(sum_{cyc}frac{a^3}{(a+b)^3}right)^2sum_{cyc}1geqleft(sum_{cyc}sqrt[3]{left(frac{a^3}{(a+b)^3}right)^2cdot1}right)^3=left(sum_{cyc}frac{a^2}{(a+b)^2}right)^3.$$
        Thus, it's enough to prove that
        $$frac{left(sumlimits_{cyc}frac{a^2}{(a+b)^2}right)^3}{3}geqfrac{9}{64}$$ or
        $$sumlimits_{cyc}frac{a^2}{(a+b)^2}geqfrac{3}{4}.$$
        Now, by C-S
        $$sumlimits_{cyc}frac{a^2}{(a+b)^2}=sumlimits_{cyc}frac{a^2(a+c)^2}{(a+b)^2(a+c)^2}geqfrac{left(sumlimits_{cyc}(a^2+ab)right)^2}{sumlimits_{cyc}(a+b)^2(a+c)^2}.$$
        Thus, it's enough to prove that
        $$4left(sumlimits_{cyc}(a^2+ab)right)^2geq3sumlimits_{cyc}(a+b)^2(a+c)^2,$$
        which is true even for all reals $a$, $b$ and $c$.



        Indeed, the last inequality is symmetric inequality by degree four,



        which says that by $uvw$ (https://math.stackexchange.com/tags/uvw/info )



        it's enough to prove the last inequality for equality case of two variables and since



        it's the homogeneous inequality by even degree, we can assume $b=c=1$, which gives
        $$(a-1)^2(a+3)^2geq0.$$
        Done!






        share|cite|improve this answer









        $endgroup$



        By Holder $$left(sum_{cyc}frac{a^3}{(a+b)^3}right)^2sum_{cyc}1geqleft(sum_{cyc}sqrt[3]{left(frac{a^3}{(a+b)^3}right)^2cdot1}right)^3=left(sum_{cyc}frac{a^2}{(a+b)^2}right)^3.$$
        Thus, it's enough to prove that
        $$frac{left(sumlimits_{cyc}frac{a^2}{(a+b)^2}right)^3}{3}geqfrac{9}{64}$$ or
        $$sumlimits_{cyc}frac{a^2}{(a+b)^2}geqfrac{3}{4}.$$
        Now, by C-S
        $$sumlimits_{cyc}frac{a^2}{(a+b)^2}=sumlimits_{cyc}frac{a^2(a+c)^2}{(a+b)^2(a+c)^2}geqfrac{left(sumlimits_{cyc}(a^2+ab)right)^2}{sumlimits_{cyc}(a+b)^2(a+c)^2}.$$
        Thus, it's enough to prove that
        $$4left(sumlimits_{cyc}(a^2+ab)right)^2geq3sumlimits_{cyc}(a+b)^2(a+c)^2,$$
        which is true even for all reals $a$, $b$ and $c$.



        Indeed, the last inequality is symmetric inequality by degree four,



        which says that by $uvw$ (https://math.stackexchange.com/tags/uvw/info )



        it's enough to prove the last inequality for equality case of two variables and since



        it's the homogeneous inequality by even degree, we can assume $b=c=1$, which gives
        $$(a-1)^2(a+3)^2geq0.$$
        Done!







        share|cite|improve this answer












        share|cite|improve this answer



        share|cite|improve this answer










        answered Dec 4 '18 at 21:35









        Michael RozenbergMichael Rozenberg

        101k1591193




        101k1591193























            5












            $begingroup$

            This is more of a comment, but I don't have the reputation. Use Lagrange multipliers. Solving, you find that the critical points occur when $xyz=1$ and $yz (1+x)^4 = xz (1+y)^4 = xy (1+z)^4$. I think the only solution is $x=y=z=1$. Clearly, it's a minimum and plugging in shows the bound.



            Added:



            We can rewrite the condition as $xyz=1$ and $frac{(1+x)^4}{x} = frac{(1+y)^4}{y} = frac{(1+z)^4}{z}$



            The function $g(x)=frac{(1+x)^4}{x}$ is decreasing from $0$ to $1/3$ and increasing from $1/3$ to $infty$. This shows that 2 of $x,y,z$ must be equal (WLOG $x$ and $y$) and $z=1/x^2$. It remains to solve $frac{(1+x)^4}{x} = frac{(1+1/x^2)^4}{1/x^2}$. This time, it's not hard to check $x=1$ is the only solution and we are done.






            share|cite|improve this answer











            $endgroup$









            • 2




              $begingroup$
              It seems to me there's enough here to post as an answer, no need to apologize.
              $endgroup$
              – David K
              Dec 4 '18 at 21:11
















            5












            $begingroup$

            This is more of a comment, but I don't have the reputation. Use Lagrange multipliers. Solving, you find that the critical points occur when $xyz=1$ and $yz (1+x)^4 = xz (1+y)^4 = xy (1+z)^4$. I think the only solution is $x=y=z=1$. Clearly, it's a minimum and plugging in shows the bound.



            Added:



            We can rewrite the condition as $xyz=1$ and $frac{(1+x)^4}{x} = frac{(1+y)^4}{y} = frac{(1+z)^4}{z}$



            The function $g(x)=frac{(1+x)^4}{x}$ is decreasing from $0$ to $1/3$ and increasing from $1/3$ to $infty$. This shows that 2 of $x,y,z$ must be equal (WLOG $x$ and $y$) and $z=1/x^2$. It remains to solve $frac{(1+x)^4}{x} = frac{(1+1/x^2)^4}{1/x^2}$. This time, it's not hard to check $x=1$ is the only solution and we are done.






            share|cite|improve this answer











            $endgroup$









            • 2




              $begingroup$
              It seems to me there's enough here to post as an answer, no need to apologize.
              $endgroup$
              – David K
              Dec 4 '18 at 21:11














            5












            5








            5





            $begingroup$

            This is more of a comment, but I don't have the reputation. Use Lagrange multipliers. Solving, you find that the critical points occur when $xyz=1$ and $yz (1+x)^4 = xz (1+y)^4 = xy (1+z)^4$. I think the only solution is $x=y=z=1$. Clearly, it's a minimum and plugging in shows the bound.



            Added:



            We can rewrite the condition as $xyz=1$ and $frac{(1+x)^4}{x} = frac{(1+y)^4}{y} = frac{(1+z)^4}{z}$



            The function $g(x)=frac{(1+x)^4}{x}$ is decreasing from $0$ to $1/3$ and increasing from $1/3$ to $infty$. This shows that 2 of $x,y,z$ must be equal (WLOG $x$ and $y$) and $z=1/x^2$. It remains to solve $frac{(1+x)^4}{x} = frac{(1+1/x^2)^4}{1/x^2}$. This time, it's not hard to check $x=1$ is the only solution and we are done.






            share|cite|improve this answer











            $endgroup$



            This is more of a comment, but I don't have the reputation. Use Lagrange multipliers. Solving, you find that the critical points occur when $xyz=1$ and $yz (1+x)^4 = xz (1+y)^4 = xy (1+z)^4$. I think the only solution is $x=y=z=1$. Clearly, it's a minimum and plugging in shows the bound.



            Added:



            We can rewrite the condition as $xyz=1$ and $frac{(1+x)^4}{x} = frac{(1+y)^4}{y} = frac{(1+z)^4}{z}$



            The function $g(x)=frac{(1+x)^4}{x}$ is decreasing from $0$ to $1/3$ and increasing from $1/3$ to $infty$. This shows that 2 of $x,y,z$ must be equal (WLOG $x$ and $y$) and $z=1/x^2$. It remains to solve $frac{(1+x)^4}{x} = frac{(1+1/x^2)^4}{1/x^2}$. This time, it's not hard to check $x=1$ is the only solution and we are done.







            share|cite|improve this answer














            share|cite|improve this answer



            share|cite|improve this answer








            edited Dec 4 '18 at 21:44

























            answered Dec 4 '18 at 20:56









            zoidbergzoidberg

            1,065113




            1,065113








            • 2




              $begingroup$
              It seems to me there's enough here to post as an answer, no need to apologize.
              $endgroup$
              – David K
              Dec 4 '18 at 21:11














            • 2




              $begingroup$
              It seems to me there's enough here to post as an answer, no need to apologize.
              $endgroup$
              – David K
              Dec 4 '18 at 21:11








            2




            2




            $begingroup$
            It seems to me there's enough here to post as an answer, no need to apologize.
            $endgroup$
            – David K
            Dec 4 '18 at 21:11




            $begingroup$
            It seems to me there's enough here to post as an answer, no need to apologize.
            $endgroup$
            – David K
            Dec 4 '18 at 21:11


















            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%2f3025819%2fprove-that-a-over-ab3b-over-bc3-c-over-ca3-geq-3-over-8%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

            Plaza Victoria

            In PowerPoint, is there a keyboard shortcut for bulleted / numbered list?

            How to put 3 figures in Latex with 2 figures side by side and 1 below these side by side images but in...