Prove that if $D^kf=0$ then $f$ is polynomial












0














Let $Esubseteqmathbb{R}^n$ an open connected set and let $f:Erightarrow mathbb{R}$ a $k+1$ times differentiable function such that $D^kf=0$. Show that $f$ is a polynomial in $x_1,...,x_n$ at a degree of at most $k$.





For any $piin{1,...,n}^n$ we're given that $frac{partial^{k+1}f}{partial x_{pi (1)}...partial x_{pi (k+1)}}=0$. Thus $frac{partial^{k}f}{partial x_{pi (1)}...partial x_{pi(k)}}=c_k$ and by induction $f=partial^0 f=x_{pi(1)}(x_{pi(2)}(...(x_{pi(k)}c_k+c_{k-1})...)+c_1)+c_0$ for some $c_iinmathbb{R}, forall 1leq i leq n$. I feel that this proof doesn't hold but I don't know where is the mistake and what is the right way. Thanks










share|cite|improve this question




















  • 3




    Taylor series might help?
    – copper.hat
    Nov 20 at 18:26










  • Your integration is false. I mean, for instance if $f(x,y)$ a function in two variables and $d^2 /dx dy f(x,y)=0$ then $d/dx f(x,y)= c(x)$ this is a function of $x$ and not just a constant. I think a proof by induction is more appropriate here.
    – Yanko
    Nov 20 at 18:28












  • You should not bring permutations into the game. The variable $k$ counting the number of differentiations is independent of the number $n$ of variables $x_i$.
    – Christian Blatter
    Nov 20 at 19:04










  • Such an $f$ is analytical and its series is finite...
    – Federico
    Nov 20 at 19:18










  • @Federico why $f$ is analytical or as you claim, entire?
    – Will M.
    Nov 20 at 19:19
















0














Let $Esubseteqmathbb{R}^n$ an open connected set and let $f:Erightarrow mathbb{R}$ a $k+1$ times differentiable function such that $D^kf=0$. Show that $f$ is a polynomial in $x_1,...,x_n$ at a degree of at most $k$.





For any $piin{1,...,n}^n$ we're given that $frac{partial^{k+1}f}{partial x_{pi (1)}...partial x_{pi (k+1)}}=0$. Thus $frac{partial^{k}f}{partial x_{pi (1)}...partial x_{pi(k)}}=c_k$ and by induction $f=partial^0 f=x_{pi(1)}(x_{pi(2)}(...(x_{pi(k)}c_k+c_{k-1})...)+c_1)+c_0$ for some $c_iinmathbb{R}, forall 1leq i leq n$. I feel that this proof doesn't hold but I don't know where is the mistake and what is the right way. Thanks










share|cite|improve this question




















  • 3




    Taylor series might help?
    – copper.hat
    Nov 20 at 18:26










  • Your integration is false. I mean, for instance if $f(x,y)$ a function in two variables and $d^2 /dx dy f(x,y)=0$ then $d/dx f(x,y)= c(x)$ this is a function of $x$ and not just a constant. I think a proof by induction is more appropriate here.
    – Yanko
    Nov 20 at 18:28












  • You should not bring permutations into the game. The variable $k$ counting the number of differentiations is independent of the number $n$ of variables $x_i$.
    – Christian Blatter
    Nov 20 at 19:04










  • Such an $f$ is analytical and its series is finite...
    – Federico
    Nov 20 at 19:18










  • @Federico why $f$ is analytical or as you claim, entire?
    – Will M.
    Nov 20 at 19:19














0












0








0







Let $Esubseteqmathbb{R}^n$ an open connected set and let $f:Erightarrow mathbb{R}$ a $k+1$ times differentiable function such that $D^kf=0$. Show that $f$ is a polynomial in $x_1,...,x_n$ at a degree of at most $k$.





For any $piin{1,...,n}^n$ we're given that $frac{partial^{k+1}f}{partial x_{pi (1)}...partial x_{pi (k+1)}}=0$. Thus $frac{partial^{k}f}{partial x_{pi (1)}...partial x_{pi(k)}}=c_k$ and by induction $f=partial^0 f=x_{pi(1)}(x_{pi(2)}(...(x_{pi(k)}c_k+c_{k-1})...)+c_1)+c_0$ for some $c_iinmathbb{R}, forall 1leq i leq n$. I feel that this proof doesn't hold but I don't know where is the mistake and what is the right way. Thanks










share|cite|improve this question















Let $Esubseteqmathbb{R}^n$ an open connected set and let $f:Erightarrow mathbb{R}$ a $k+1$ times differentiable function such that $D^kf=0$. Show that $f$ is a polynomial in $x_1,...,x_n$ at a degree of at most $k$.





For any $piin{1,...,n}^n$ we're given that $frac{partial^{k+1}f}{partial x_{pi (1)}...partial x_{pi (k+1)}}=0$. Thus $frac{partial^{k}f}{partial x_{pi (1)}...partial x_{pi(k)}}=c_k$ and by induction $f=partial^0 f=x_{pi(1)}(x_{pi(2)}(...(x_{pi(k)}c_k+c_{k-1})...)+c_1)+c_0$ for some $c_iinmathbb{R}, forall 1leq i leq n$. I feel that this proof doesn't hold but I don't know where is the mistake and what is the right way. Thanks







calculus derivatives partial-derivative






share|cite|improve this question















share|cite|improve this question













share|cite|improve this question




share|cite|improve this question








edited Nov 20 at 18:26

























asked Nov 20 at 18:24









J. Doe

976




976








  • 3




    Taylor series might help?
    – copper.hat
    Nov 20 at 18:26










  • Your integration is false. I mean, for instance if $f(x,y)$ a function in two variables and $d^2 /dx dy f(x,y)=0$ then $d/dx f(x,y)= c(x)$ this is a function of $x$ and not just a constant. I think a proof by induction is more appropriate here.
    – Yanko
    Nov 20 at 18:28












  • You should not bring permutations into the game. The variable $k$ counting the number of differentiations is independent of the number $n$ of variables $x_i$.
    – Christian Blatter
    Nov 20 at 19:04










  • Such an $f$ is analytical and its series is finite...
    – Federico
    Nov 20 at 19:18










  • @Federico why $f$ is analytical or as you claim, entire?
    – Will M.
    Nov 20 at 19:19














  • 3




    Taylor series might help?
    – copper.hat
    Nov 20 at 18:26










  • Your integration is false. I mean, for instance if $f(x,y)$ a function in two variables and $d^2 /dx dy f(x,y)=0$ then $d/dx f(x,y)= c(x)$ this is a function of $x$ and not just a constant. I think a proof by induction is more appropriate here.
    – Yanko
    Nov 20 at 18:28












  • You should not bring permutations into the game. The variable $k$ counting the number of differentiations is independent of the number $n$ of variables $x_i$.
    – Christian Blatter
    Nov 20 at 19:04










  • Such an $f$ is analytical and its series is finite...
    – Federico
    Nov 20 at 19:18










  • @Federico why $f$ is analytical or as you claim, entire?
    – Will M.
    Nov 20 at 19:19








3




3




Taylor series might help?
– copper.hat
Nov 20 at 18:26




Taylor series might help?
– copper.hat
Nov 20 at 18:26












Your integration is false. I mean, for instance if $f(x,y)$ a function in two variables and $d^2 /dx dy f(x,y)=0$ then $d/dx f(x,y)= c(x)$ this is a function of $x$ and not just a constant. I think a proof by induction is more appropriate here.
– Yanko
Nov 20 at 18:28






Your integration is false. I mean, for instance if $f(x,y)$ a function in two variables and $d^2 /dx dy f(x,y)=0$ then $d/dx f(x,y)= c(x)$ this is a function of $x$ and not just a constant. I think a proof by induction is more appropriate here.
– Yanko
Nov 20 at 18:28














You should not bring permutations into the game. The variable $k$ counting the number of differentiations is independent of the number $n$ of variables $x_i$.
– Christian Blatter
Nov 20 at 19:04




You should not bring permutations into the game. The variable $k$ counting the number of differentiations is independent of the number $n$ of variables $x_i$.
– Christian Blatter
Nov 20 at 19:04












Such an $f$ is analytical and its series is finite...
– Federico
Nov 20 at 19:18




Such an $f$ is analytical and its series is finite...
– Federico
Nov 20 at 19:18












@Federico why $f$ is analytical or as you claim, entire?
– Will M.
Nov 20 at 19:19




@Federico why $f$ is analytical or as you claim, entire?
– Will M.
Nov 20 at 19:19










3 Answers
3






active

oldest

votes


















0














Abridged solution. Proceed by induction on $k.$ The case $k = 0$ is very standard; consider a "base point" $a in E$ and the set $mathrm{X}$ of points in $x in mathrm{E}$ such that $f(x)=f(a).$ Then $mathrm{X}$ is closed in $mathrm{E}$ as $f$ is continuous. Now, $mathrm{X}$ is also open in $mathrm{E}$ for if $x in mathrm{X}$ then there is a small enough ball centred at $x$ contained in $mathrm{E}$ and for such ball the mean value theorem (applied to the line segment between $x$ and another point in the ball) gives that $f$ is constant in such ball, so such ball is subset of $mathrm{X}$ and then $mathrm{X}$ is open as claimed. Being $mathrm{E}$ connected, $mathrm{X} = mathrm{E}$ and the base case $k = 0$ is proved.



For the general case, assume $f$ satisfies $f^{(k+1)}=0$ on $mathrm{E}.$ The case $k = 0$ shows that $f^{(k)}$ is a constant $k$-linear function on $mathrm{E},$ call it $M.$ We show that $varphi:x mapsto f(x) - dfrac{1}{k!}M(x, ldots, x)$ satisfies $varphi^{(k)}=0$ on $mathrm{E}$ and hence the result. By standard results about derivatives of multilinear functions and the chain rule, we get at once that $varphi^{(k)}=f^{(k)}-dfrac{1}{k!} k!M = 0.$ Q.E.D.






share|cite|improve this answer





























    0














    Note that $f$ is smooth since if $D^k f = 0$ then $D^l f = 0$ for $ l ge k$.



    Fix some $x_0 in E$, and suppose $C$ is an open convex set containing $x_0$ and contained in $E$. Then Taylor's theorem shows that if $y in C$ then we
    have $f(y) = sum_{|alpha| le k-1} {1 over alpha !} D^alpha f(x_0) (y-x_0)^alpha$.



    Let $p(y) = sum_{|alpha| le k-1} {1 over alpha !} D^alpha f(x_0) (y-x_0)^alpha$ be the polynomial, note that $p$ is defined everywhere and has maximum degree $k-1$. We wish to show that
    $f=p$ on $E$.



    As an aside, note that if $q$ is a polynomial and $q(x) = 0$ for $x$ in some open set then $q=0$.



    Pick some $x_nin E$ and note that since $E$ is open and connected there are $x_k in E$ such that the straight line curve $x_0 to x_1 to cdots to x_n$ is contained
    in $E$.



    Since the segment $[x_0,x_1] subset E$, there is some open convex set $C_1$ such that
    $[x_0,x_1] subset C_1 subset E$ and we have $f=p$ on $C_1$.



    Now let $p'$ be the Taylor expansion based at $x_1$. As above, there is some open
    convex set with $[x_1,x_2] subset C_2 subset E$ and we have $f=p'$ on $C_2$.



    Since $x_1 in C_1 cap C_2$, we see that $p=p'$ on an open set, and hence $p'=p$ everywhere.



    It is clear that we can repeat this process to see that $f(x_n) = p(x_n)$.



    Since $x_n in E$ was arbitrary, we see that $f=p$ on $E$.






    share|cite|improve this answer





























      -2














      Another alternative proof (which I guess is worng as well because I don't use the fact that $E$ is open and connected):



      From Taylor's formula we know that
      $$
      \ f(x_1,...,x_n)=sum_{i_1=1}^infty ...sum_{i_n=1}^infty frac{(x_1-a_1)^{i_1}cdot...cdot(x_n-a_n)^{i_n}}{i_1!cdot...cdot i_n!}cdotfrac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a_1,...,a_n)
      $$

      Let $0leq i_1,...,i_nleq k$. If $sum_{j=1}^n i_j>k$ then $frac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a)=0$. Otherwise, $frac{(x_1-a_1)^{i_1}cdot...cdot(x_n-a_n)^{i_n}}{i_1!cdot...cdot i_n!}$ is polynom of a degree at most $k$, and $frac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a_1,...,a_n)$ is a constant linear tranformation. Thus $f(x_1,...,x_n)$ is a polynom of degree at most $k$.






      share|cite|improve this answer





















      • Why $f$ is analytic?
        – Will M.
        Nov 20 at 19:18










      • I don't find the answre, why is it really? @WillM.
        – J. Doe
        Nov 20 at 19:32










      • I can only prove it is by well, showing it is a polynomial. This phenomenon of "a function is a polynomial provided one of its higher derivatives is zero identically" holds true in Banach spaces. As the proof I wrote shows.
        – Will M.
        Nov 20 at 19:56













      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%2f3006708%2fprove-that-if-dkf-0-then-f-is-polynomial%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









      0














      Abridged solution. Proceed by induction on $k.$ The case $k = 0$ is very standard; consider a "base point" $a in E$ and the set $mathrm{X}$ of points in $x in mathrm{E}$ such that $f(x)=f(a).$ Then $mathrm{X}$ is closed in $mathrm{E}$ as $f$ is continuous. Now, $mathrm{X}$ is also open in $mathrm{E}$ for if $x in mathrm{X}$ then there is a small enough ball centred at $x$ contained in $mathrm{E}$ and for such ball the mean value theorem (applied to the line segment between $x$ and another point in the ball) gives that $f$ is constant in such ball, so such ball is subset of $mathrm{X}$ and then $mathrm{X}$ is open as claimed. Being $mathrm{E}$ connected, $mathrm{X} = mathrm{E}$ and the base case $k = 0$ is proved.



      For the general case, assume $f$ satisfies $f^{(k+1)}=0$ on $mathrm{E}.$ The case $k = 0$ shows that $f^{(k)}$ is a constant $k$-linear function on $mathrm{E},$ call it $M.$ We show that $varphi:x mapsto f(x) - dfrac{1}{k!}M(x, ldots, x)$ satisfies $varphi^{(k)}=0$ on $mathrm{E}$ and hence the result. By standard results about derivatives of multilinear functions and the chain rule, we get at once that $varphi^{(k)}=f^{(k)}-dfrac{1}{k!} k!M = 0.$ Q.E.D.






      share|cite|improve this answer


























        0














        Abridged solution. Proceed by induction on $k.$ The case $k = 0$ is very standard; consider a "base point" $a in E$ and the set $mathrm{X}$ of points in $x in mathrm{E}$ such that $f(x)=f(a).$ Then $mathrm{X}$ is closed in $mathrm{E}$ as $f$ is continuous. Now, $mathrm{X}$ is also open in $mathrm{E}$ for if $x in mathrm{X}$ then there is a small enough ball centred at $x$ contained in $mathrm{E}$ and for such ball the mean value theorem (applied to the line segment between $x$ and another point in the ball) gives that $f$ is constant in such ball, so such ball is subset of $mathrm{X}$ and then $mathrm{X}$ is open as claimed. Being $mathrm{E}$ connected, $mathrm{X} = mathrm{E}$ and the base case $k = 0$ is proved.



        For the general case, assume $f$ satisfies $f^{(k+1)}=0$ on $mathrm{E}.$ The case $k = 0$ shows that $f^{(k)}$ is a constant $k$-linear function on $mathrm{E},$ call it $M.$ We show that $varphi:x mapsto f(x) - dfrac{1}{k!}M(x, ldots, x)$ satisfies $varphi^{(k)}=0$ on $mathrm{E}$ and hence the result. By standard results about derivatives of multilinear functions and the chain rule, we get at once that $varphi^{(k)}=f^{(k)}-dfrac{1}{k!} k!M = 0.$ Q.E.D.






        share|cite|improve this answer
























          0












          0








          0






          Abridged solution. Proceed by induction on $k.$ The case $k = 0$ is very standard; consider a "base point" $a in E$ and the set $mathrm{X}$ of points in $x in mathrm{E}$ such that $f(x)=f(a).$ Then $mathrm{X}$ is closed in $mathrm{E}$ as $f$ is continuous. Now, $mathrm{X}$ is also open in $mathrm{E}$ for if $x in mathrm{X}$ then there is a small enough ball centred at $x$ contained in $mathrm{E}$ and for such ball the mean value theorem (applied to the line segment between $x$ and another point in the ball) gives that $f$ is constant in such ball, so such ball is subset of $mathrm{X}$ and then $mathrm{X}$ is open as claimed. Being $mathrm{E}$ connected, $mathrm{X} = mathrm{E}$ and the base case $k = 0$ is proved.



          For the general case, assume $f$ satisfies $f^{(k+1)}=0$ on $mathrm{E}.$ The case $k = 0$ shows that $f^{(k)}$ is a constant $k$-linear function on $mathrm{E},$ call it $M.$ We show that $varphi:x mapsto f(x) - dfrac{1}{k!}M(x, ldots, x)$ satisfies $varphi^{(k)}=0$ on $mathrm{E}$ and hence the result. By standard results about derivatives of multilinear functions and the chain rule, we get at once that $varphi^{(k)}=f^{(k)}-dfrac{1}{k!} k!M = 0.$ Q.E.D.






          share|cite|improve this answer












          Abridged solution. Proceed by induction on $k.$ The case $k = 0$ is very standard; consider a "base point" $a in E$ and the set $mathrm{X}$ of points in $x in mathrm{E}$ such that $f(x)=f(a).$ Then $mathrm{X}$ is closed in $mathrm{E}$ as $f$ is continuous. Now, $mathrm{X}$ is also open in $mathrm{E}$ for if $x in mathrm{X}$ then there is a small enough ball centred at $x$ contained in $mathrm{E}$ and for such ball the mean value theorem (applied to the line segment between $x$ and another point in the ball) gives that $f$ is constant in such ball, so such ball is subset of $mathrm{X}$ and then $mathrm{X}$ is open as claimed. Being $mathrm{E}$ connected, $mathrm{X} = mathrm{E}$ and the base case $k = 0$ is proved.



          For the general case, assume $f$ satisfies $f^{(k+1)}=0$ on $mathrm{E}.$ The case $k = 0$ shows that $f^{(k)}$ is a constant $k$-linear function on $mathrm{E},$ call it $M.$ We show that $varphi:x mapsto f(x) - dfrac{1}{k!}M(x, ldots, x)$ satisfies $varphi^{(k)}=0$ on $mathrm{E}$ and hence the result. By standard results about derivatives of multilinear functions and the chain rule, we get at once that $varphi^{(k)}=f^{(k)}-dfrac{1}{k!} k!M = 0.$ Q.E.D.







          share|cite|improve this answer












          share|cite|improve this answer



          share|cite|improve this answer










          answered Nov 20 at 19:33









          Will M.

          2,377314




          2,377314























              0














              Note that $f$ is smooth since if $D^k f = 0$ then $D^l f = 0$ for $ l ge k$.



              Fix some $x_0 in E$, and suppose $C$ is an open convex set containing $x_0$ and contained in $E$. Then Taylor's theorem shows that if $y in C$ then we
              have $f(y) = sum_{|alpha| le k-1} {1 over alpha !} D^alpha f(x_0) (y-x_0)^alpha$.



              Let $p(y) = sum_{|alpha| le k-1} {1 over alpha !} D^alpha f(x_0) (y-x_0)^alpha$ be the polynomial, note that $p$ is defined everywhere and has maximum degree $k-1$. We wish to show that
              $f=p$ on $E$.



              As an aside, note that if $q$ is a polynomial and $q(x) = 0$ for $x$ in some open set then $q=0$.



              Pick some $x_nin E$ and note that since $E$ is open and connected there are $x_k in E$ such that the straight line curve $x_0 to x_1 to cdots to x_n$ is contained
              in $E$.



              Since the segment $[x_0,x_1] subset E$, there is some open convex set $C_1$ such that
              $[x_0,x_1] subset C_1 subset E$ and we have $f=p$ on $C_1$.



              Now let $p'$ be the Taylor expansion based at $x_1$. As above, there is some open
              convex set with $[x_1,x_2] subset C_2 subset E$ and we have $f=p'$ on $C_2$.



              Since $x_1 in C_1 cap C_2$, we see that $p=p'$ on an open set, and hence $p'=p$ everywhere.



              It is clear that we can repeat this process to see that $f(x_n) = p(x_n)$.



              Since $x_n in E$ was arbitrary, we see that $f=p$ on $E$.






              share|cite|improve this answer


























                0














                Note that $f$ is smooth since if $D^k f = 0$ then $D^l f = 0$ for $ l ge k$.



                Fix some $x_0 in E$, and suppose $C$ is an open convex set containing $x_0$ and contained in $E$. Then Taylor's theorem shows that if $y in C$ then we
                have $f(y) = sum_{|alpha| le k-1} {1 over alpha !} D^alpha f(x_0) (y-x_0)^alpha$.



                Let $p(y) = sum_{|alpha| le k-1} {1 over alpha !} D^alpha f(x_0) (y-x_0)^alpha$ be the polynomial, note that $p$ is defined everywhere and has maximum degree $k-1$. We wish to show that
                $f=p$ on $E$.



                As an aside, note that if $q$ is a polynomial and $q(x) = 0$ for $x$ in some open set then $q=0$.



                Pick some $x_nin E$ and note that since $E$ is open and connected there are $x_k in E$ such that the straight line curve $x_0 to x_1 to cdots to x_n$ is contained
                in $E$.



                Since the segment $[x_0,x_1] subset E$, there is some open convex set $C_1$ such that
                $[x_0,x_1] subset C_1 subset E$ and we have $f=p$ on $C_1$.



                Now let $p'$ be the Taylor expansion based at $x_1$. As above, there is some open
                convex set with $[x_1,x_2] subset C_2 subset E$ and we have $f=p'$ on $C_2$.



                Since $x_1 in C_1 cap C_2$, we see that $p=p'$ on an open set, and hence $p'=p$ everywhere.



                It is clear that we can repeat this process to see that $f(x_n) = p(x_n)$.



                Since $x_n in E$ was arbitrary, we see that $f=p$ on $E$.






                share|cite|improve this answer
























                  0












                  0








                  0






                  Note that $f$ is smooth since if $D^k f = 0$ then $D^l f = 0$ for $ l ge k$.



                  Fix some $x_0 in E$, and suppose $C$ is an open convex set containing $x_0$ and contained in $E$. Then Taylor's theorem shows that if $y in C$ then we
                  have $f(y) = sum_{|alpha| le k-1} {1 over alpha !} D^alpha f(x_0) (y-x_0)^alpha$.



                  Let $p(y) = sum_{|alpha| le k-1} {1 over alpha !} D^alpha f(x_0) (y-x_0)^alpha$ be the polynomial, note that $p$ is defined everywhere and has maximum degree $k-1$. We wish to show that
                  $f=p$ on $E$.



                  As an aside, note that if $q$ is a polynomial and $q(x) = 0$ for $x$ in some open set then $q=0$.



                  Pick some $x_nin E$ and note that since $E$ is open and connected there are $x_k in E$ such that the straight line curve $x_0 to x_1 to cdots to x_n$ is contained
                  in $E$.



                  Since the segment $[x_0,x_1] subset E$, there is some open convex set $C_1$ such that
                  $[x_0,x_1] subset C_1 subset E$ and we have $f=p$ on $C_1$.



                  Now let $p'$ be the Taylor expansion based at $x_1$. As above, there is some open
                  convex set with $[x_1,x_2] subset C_2 subset E$ and we have $f=p'$ on $C_2$.



                  Since $x_1 in C_1 cap C_2$, we see that $p=p'$ on an open set, and hence $p'=p$ everywhere.



                  It is clear that we can repeat this process to see that $f(x_n) = p(x_n)$.



                  Since $x_n in E$ was arbitrary, we see that $f=p$ on $E$.






                  share|cite|improve this answer












                  Note that $f$ is smooth since if $D^k f = 0$ then $D^l f = 0$ for $ l ge k$.



                  Fix some $x_0 in E$, and suppose $C$ is an open convex set containing $x_0$ and contained in $E$. Then Taylor's theorem shows that if $y in C$ then we
                  have $f(y) = sum_{|alpha| le k-1} {1 over alpha !} D^alpha f(x_0) (y-x_0)^alpha$.



                  Let $p(y) = sum_{|alpha| le k-1} {1 over alpha !} D^alpha f(x_0) (y-x_0)^alpha$ be the polynomial, note that $p$ is defined everywhere and has maximum degree $k-1$. We wish to show that
                  $f=p$ on $E$.



                  As an aside, note that if $q$ is a polynomial and $q(x) = 0$ for $x$ in some open set then $q=0$.



                  Pick some $x_nin E$ and note that since $E$ is open and connected there are $x_k in E$ such that the straight line curve $x_0 to x_1 to cdots to x_n$ is contained
                  in $E$.



                  Since the segment $[x_0,x_1] subset E$, there is some open convex set $C_1$ such that
                  $[x_0,x_1] subset C_1 subset E$ and we have $f=p$ on $C_1$.



                  Now let $p'$ be the Taylor expansion based at $x_1$. As above, there is some open
                  convex set with $[x_1,x_2] subset C_2 subset E$ and we have $f=p'$ on $C_2$.



                  Since $x_1 in C_1 cap C_2$, we see that $p=p'$ on an open set, and hence $p'=p$ everywhere.



                  It is clear that we can repeat this process to see that $f(x_n) = p(x_n)$.



                  Since $x_n in E$ was arbitrary, we see that $f=p$ on $E$.







                  share|cite|improve this answer












                  share|cite|improve this answer



                  share|cite|improve this answer










                  answered Nov 24 at 4:21









                  copper.hat

                  126k559159




                  126k559159























                      -2














                      Another alternative proof (which I guess is worng as well because I don't use the fact that $E$ is open and connected):



                      From Taylor's formula we know that
                      $$
                      \ f(x_1,...,x_n)=sum_{i_1=1}^infty ...sum_{i_n=1}^infty frac{(x_1-a_1)^{i_1}cdot...cdot(x_n-a_n)^{i_n}}{i_1!cdot...cdot i_n!}cdotfrac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a_1,...,a_n)
                      $$

                      Let $0leq i_1,...,i_nleq k$. If $sum_{j=1}^n i_j>k$ then $frac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a)=0$. Otherwise, $frac{(x_1-a_1)^{i_1}cdot...cdot(x_n-a_n)^{i_n}}{i_1!cdot...cdot i_n!}$ is polynom of a degree at most $k$, and $frac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a_1,...,a_n)$ is a constant linear tranformation. Thus $f(x_1,...,x_n)$ is a polynom of degree at most $k$.






                      share|cite|improve this answer





















                      • Why $f$ is analytic?
                        – Will M.
                        Nov 20 at 19:18










                      • I don't find the answre, why is it really? @WillM.
                        – J. Doe
                        Nov 20 at 19:32










                      • I can only prove it is by well, showing it is a polynomial. This phenomenon of "a function is a polynomial provided one of its higher derivatives is zero identically" holds true in Banach spaces. As the proof I wrote shows.
                        – Will M.
                        Nov 20 at 19:56


















                      -2














                      Another alternative proof (which I guess is worng as well because I don't use the fact that $E$ is open and connected):



                      From Taylor's formula we know that
                      $$
                      \ f(x_1,...,x_n)=sum_{i_1=1}^infty ...sum_{i_n=1}^infty frac{(x_1-a_1)^{i_1}cdot...cdot(x_n-a_n)^{i_n}}{i_1!cdot...cdot i_n!}cdotfrac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a_1,...,a_n)
                      $$

                      Let $0leq i_1,...,i_nleq k$. If $sum_{j=1}^n i_j>k$ then $frac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a)=0$. Otherwise, $frac{(x_1-a_1)^{i_1}cdot...cdot(x_n-a_n)^{i_n}}{i_1!cdot...cdot i_n!}$ is polynom of a degree at most $k$, and $frac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a_1,...,a_n)$ is a constant linear tranformation. Thus $f(x_1,...,x_n)$ is a polynom of degree at most $k$.






                      share|cite|improve this answer





















                      • Why $f$ is analytic?
                        – Will M.
                        Nov 20 at 19:18










                      • I don't find the answre, why is it really? @WillM.
                        – J. Doe
                        Nov 20 at 19:32










                      • I can only prove it is by well, showing it is a polynomial. This phenomenon of "a function is a polynomial provided one of its higher derivatives is zero identically" holds true in Banach spaces. As the proof I wrote shows.
                        – Will M.
                        Nov 20 at 19:56
















                      -2












                      -2








                      -2






                      Another alternative proof (which I guess is worng as well because I don't use the fact that $E$ is open and connected):



                      From Taylor's formula we know that
                      $$
                      \ f(x_1,...,x_n)=sum_{i_1=1}^infty ...sum_{i_n=1}^infty frac{(x_1-a_1)^{i_1}cdot...cdot(x_n-a_n)^{i_n}}{i_1!cdot...cdot i_n!}cdotfrac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a_1,...,a_n)
                      $$

                      Let $0leq i_1,...,i_nleq k$. If $sum_{j=1}^n i_j>k$ then $frac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a)=0$. Otherwise, $frac{(x_1-a_1)^{i_1}cdot...cdot(x_n-a_n)^{i_n}}{i_1!cdot...cdot i_n!}$ is polynom of a degree at most $k$, and $frac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a_1,...,a_n)$ is a constant linear tranformation. Thus $f(x_1,...,x_n)$ is a polynom of degree at most $k$.






                      share|cite|improve this answer












                      Another alternative proof (which I guess is worng as well because I don't use the fact that $E$ is open and connected):



                      From Taylor's formula we know that
                      $$
                      \ f(x_1,...,x_n)=sum_{i_1=1}^infty ...sum_{i_n=1}^infty frac{(x_1-a_1)^{i_1}cdot...cdot(x_n-a_n)^{i_n}}{i_1!cdot...cdot i_n!}cdotfrac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a_1,...,a_n)
                      $$

                      Let $0leq i_1,...,i_nleq k$. If $sum_{j=1}^n i_j>k$ then $frac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a)=0$. Otherwise, $frac{(x_1-a_1)^{i_1}cdot...cdot(x_n-a_n)^{i_n}}{i_1!cdot...cdot i_n!}$ is polynom of a degree at most $k$, and $frac{partial ^{i_1+...+i_n}f}{partial x_1^{i_1}...partial x_n^{i_n}}(a_1,...,a_n)$ is a constant linear tranformation. Thus $f(x_1,...,x_n)$ is a polynom of degree at most $k$.







                      share|cite|improve this answer












                      share|cite|improve this answer



                      share|cite|improve this answer










                      answered Nov 20 at 19:11









                      J. Doe

                      976




                      976












                      • Why $f$ is analytic?
                        – Will M.
                        Nov 20 at 19:18










                      • I don't find the answre, why is it really? @WillM.
                        – J. Doe
                        Nov 20 at 19:32










                      • I can only prove it is by well, showing it is a polynomial. This phenomenon of "a function is a polynomial provided one of its higher derivatives is zero identically" holds true in Banach spaces. As the proof I wrote shows.
                        – Will M.
                        Nov 20 at 19:56




















                      • Why $f$ is analytic?
                        – Will M.
                        Nov 20 at 19:18










                      • I don't find the answre, why is it really? @WillM.
                        – J. Doe
                        Nov 20 at 19:32










                      • I can only prove it is by well, showing it is a polynomial. This phenomenon of "a function is a polynomial provided one of its higher derivatives is zero identically" holds true in Banach spaces. As the proof I wrote shows.
                        – Will M.
                        Nov 20 at 19:56


















                      Why $f$ is analytic?
                      – Will M.
                      Nov 20 at 19:18




                      Why $f$ is analytic?
                      – Will M.
                      Nov 20 at 19:18












                      I don't find the answre, why is it really? @WillM.
                      – J. Doe
                      Nov 20 at 19:32




                      I don't find the answre, why is it really? @WillM.
                      – J. Doe
                      Nov 20 at 19:32












                      I can only prove it is by well, showing it is a polynomial. This phenomenon of "a function is a polynomial provided one of its higher derivatives is zero identically" holds true in Banach spaces. As the proof I wrote shows.
                      – Will M.
                      Nov 20 at 19:56






                      I can only prove it is by well, showing it is a polynomial. This phenomenon of "a function is a polynomial provided one of its higher derivatives is zero identically" holds true in Banach spaces. As the proof I wrote shows.
                      – Will M.
                      Nov 20 at 19:56




















                      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.





                      Some of your past answers have not been well-received, and you're in danger of being blocked from answering.


                      Please pay close attention to the following guidance:


                      • 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.


                      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%2f3006708%2fprove-that-if-dkf-0-then-f-is-polynomial%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

                      Biblatex bibliography style without URLs when DOI exists (in Overleaf with Zotero bibliography)

                      ComboBox Display Member on multiple fields

                      Is it possible to collect Nectar points via Trainline?