Show this function on sequences is cyclical.

The name of the pictureThe name of the pictureThe name of the pictureClash Royale CLAN TAG#URR8PPP











up vote
10
down vote

favorite
2












Given a sequence $f$ define a new sequence $phi f$ as follows:



$$ phi f(n):=textcard kinmathbbN_leqnmid f(k)=f(n)$$



where $textcard$ is the set cardinality function. In other words $phi f(n)$ counts up how often $f(n)=f(k)$ for $kleq n$. Also, we can iteratively apply $phi$ to $f$. Here is an example:



$$beginarray \
f & = & (a,&b,&a,&c,&b,&d,&a,&c,&a,&b,&...) \
phi f & = & (1,&1,&2,&1,&2,&1,&3,&2,&4,&3,&...) \
phi^2 f & = & (1,&2,&1,&3,&2,&4,&1,&3,&1,&2,&...) \
phi^3 f & = & (1,&1,&2,&1,&2,&1,&3,&2,&4,&3,&...) \
endarray$$



Notice that $phi f=phi^3 f$, which via induction implies $phi^nf=phi^n+2f$ for all $ninmathbbN$




Conjecture : If $f$ is any sequence then $phi^n f = phi^n+2 f$ for all $ninmathbbN$.








share|cite|improve this question


























    up vote
    10
    down vote

    favorite
    2












    Given a sequence $f$ define a new sequence $phi f$ as follows:



    $$ phi f(n):=textcard kinmathbbN_leqnmid f(k)=f(n)$$



    where $textcard$ is the set cardinality function. In other words $phi f(n)$ counts up how often $f(n)=f(k)$ for $kleq n$. Also, we can iteratively apply $phi$ to $f$. Here is an example:



    $$beginarray \
    f & = & (a,&b,&a,&c,&b,&d,&a,&c,&a,&b,&...) \
    phi f & = & (1,&1,&2,&1,&2,&1,&3,&2,&4,&3,&...) \
    phi^2 f & = & (1,&2,&1,&3,&2,&4,&1,&3,&1,&2,&...) \
    phi^3 f & = & (1,&1,&2,&1,&2,&1,&3,&2,&4,&3,&...) \
    endarray$$



    Notice that $phi f=phi^3 f$, which via induction implies $phi^nf=phi^n+2f$ for all $ninmathbbN$




    Conjecture : If $f$ is any sequence then $phi^n f = phi^n+2 f$ for all $ninmathbbN$.








    share|cite|improve this question
























      up vote
      10
      down vote

      favorite
      2









      up vote
      10
      down vote

      favorite
      2






      2





      Given a sequence $f$ define a new sequence $phi f$ as follows:



      $$ phi f(n):=textcard kinmathbbN_leqnmid f(k)=f(n)$$



      where $textcard$ is the set cardinality function. In other words $phi f(n)$ counts up how often $f(n)=f(k)$ for $kleq n$. Also, we can iteratively apply $phi$ to $f$. Here is an example:



      $$beginarray \
      f & = & (a,&b,&a,&c,&b,&d,&a,&c,&a,&b,&...) \
      phi f & = & (1,&1,&2,&1,&2,&1,&3,&2,&4,&3,&...) \
      phi^2 f & = & (1,&2,&1,&3,&2,&4,&1,&3,&1,&2,&...) \
      phi^3 f & = & (1,&1,&2,&1,&2,&1,&3,&2,&4,&3,&...) \
      endarray$$



      Notice that $phi f=phi^3 f$, which via induction implies $phi^nf=phi^n+2f$ for all $ninmathbbN$




      Conjecture : If $f$ is any sequence then $phi^n f = phi^n+2 f$ for all $ninmathbbN$.








      share|cite|improve this question














      Given a sequence $f$ define a new sequence $phi f$ as follows:



      $$ phi f(n):=textcard kinmathbbN_leqnmid f(k)=f(n)$$



      where $textcard$ is the set cardinality function. In other words $phi f(n)$ counts up how often $f(n)=f(k)$ for $kleq n$. Also, we can iteratively apply $phi$ to $f$. Here is an example:



      $$beginarray \
      f & = & (a,&b,&a,&c,&b,&d,&a,&c,&a,&b,&...) \
      phi f & = & (1,&1,&2,&1,&2,&1,&3,&2,&4,&3,&...) \
      phi^2 f & = & (1,&2,&1,&3,&2,&4,&1,&3,&1,&2,&...) \
      phi^3 f & = & (1,&1,&2,&1,&2,&1,&3,&2,&4,&3,&...) \
      endarray$$



      Notice that $phi f=phi^3 f$, which via induction implies $phi^nf=phi^n+2f$ for all $ninmathbbN$




      Conjecture : If $f$ is any sequence then $phi^n f = phi^n+2 f$ for all $ninmathbbN$.










      share|cite|improve this question













      share|cite|improve this question




      share|cite|improve this question








      edited 2 days ago

























      asked Aug 26 at 20:51









      M. Nestor

      4519




      4519




















          2 Answers
          2






          active

          oldest

          votes

















          up vote
          4
          down vote



          accepted










          Very nice observation.

          As you note, for the conjecture it suffices to prove $phi^3=phi$.



          Call sequences $a=(a_1,a_2,dots)$ and $b=(b_1,b_2,dots)$ similar, if $phi a=phi b$, i.e. for each index $n$, $ a_n$ occurs the same times among $a_1,dots,a_n $ as $ b_n$ among $b_1,dots,b_n$.

          For example, the sequences $(1,1,2,2,3,3,1), (1,1,2,2,3,3,2), (1,1,2,2,3,3,3)$ are similar.



          We have to show that $phi^2 a$ is similar to $a$ for an arbitrary sequence $a$.



          Assume it holds for $n$ long sequences, and take $a_n+1$ in $a=(a_1,dots,a_n+1)$.



          • It is either a 'new element' ($phi a(n+1)=1$), in which case the number of $1$'s that denotes the new elements, is increased by one in $phi a$, implying $phi^2 a(n+1)=|a_1,dots,a_n+1|$ which is a new element in the sequence $phi^2a$.

          • Or, it is a repeating element, and $s:=phi a(n+1)=|ile n+1:a_i=a_n+1|$.

            This $s$ occurs in $phi a$ for exactly $q$ times, where $q$ is the number of $a_i$'s occuring exactly $s$ times in $a$. (If $s$ is new, then $q=1$, and $a_n+1$ can be replaced by any of these $a_i$'s to get a similar sequence.)

            On one hand, it means $phi^2 a(n+1)=q$.

            On the other hand, $q$ already occured in $phi^2 a(1,dots,n)$ exactly $s-1=|ile n:a_i=a_n+1|$ times.


          Another approach



          Call a sequence $a$ of positive integers a regular sequence, if for all $n$,



          1. $ 1le a_nlemax_i<na_i+1$

          2. In every initial segment of $a$, the number of $x$'s is at least the number of $y$'s whenever $xle y$.

          Note that 2. can be rephrased as: if $a_n=a_i$ for some $i<n$, then $a_n$ is minimal among those sequence elements which occur exactly $(phi a)_n-1$ times in $(a_1,dots,a_n-1)$

          [because $a_n$ in $(a_1,dots,a_n)$
          has one more occurances than any of them].



          For example, $(1,2,2)$ and $(1,2,3,2)$ are not regular, while $(1,2,1)$ and $(1,2,3,1)$ are regular.



          Proposition.

          $ $ a) $ phi a$ is regular for any sequence $a$.

          $ $ b) $ $If $a$ is a regular sequence, then $phi^2 a=a$.






          share|cite|improve this answer






















          • Currently considering "level sets" of the sequence. If the original sequence $f$ has domain $D$ then for $xin D$ I let $E_x=ninmathbbNmid f(n)=x$. If you list those elements like $E_x=e_1,e_2,...$ then $f(e_n)=n$, and $phi f(e_n)$ is a constant, and furthermore $E_x$ is the only set which is mapped to this constant. Thus, $phi^2 f(e_n)=f(e_n)=n$ for all $n$.
            – M. Nestor
            Aug 26 at 23:42











          • Yes, level sets definitely play some role here around.
            – Berci
            Aug 26 at 23:46










          • Hmm.. You still have to finish the proof (of the last sentence).
            – Berci
            Aug 27 at 0:32










          • I think I can prove the case for the level sets, essentially that $(1,2,3,...)mapsto(1,1,1,...)mapsto(1,2,3,...)$, and then a larger argument piecing them together for an arbitrary sequence.
            – M. Nestor
            Aug 27 at 0:52










          • You're allowed to post your own answer. What is that argument?
            – Berci
            Aug 27 at 6:11


















          up vote
          1
          down vote













          Solution in progress



          For a sequence $f$ define



          1. the pre-image of $f$ denoted $f^-1(a):=ninmathbbNmid f(n)=a$.


          2. a truncation of $f$ at $t$, denoted $f_leq t$, is the sequence with the domain restricted to $ninmathbbNmid nleq t$.


          Then by $f_leq t^-1(n)$ denote $kinmathbbNmid kleq t,f(k)=f(n)$. Here is a new definition for $phi$:




          Definition 1: For any sequence $f$, define $phi f(n) := textcard f_leq n^-1(n) $




          The following definition and conjecture describes the sequences that $phi$ will produce:




          Definition 2: A sequence $f$ with range $mathbbN$ is regular if for any truncation at $tinmathbbN$, $$textcard f_leq
          t^-1(1)geqtextcard f_leq t^-1(2)geqtextcard f_leq
          t^-1(3)geq...$$




          That is, for any truncation of the sequence $f$, the number of $1$s is $geq$ the number of $2$s, which is $geq$ the number of $3$s and so on, which ensures the sequence is counting properly. I am nearly certain this definition is equivalent to those in comments.




          Conjecture: If $f$ is any sequence then $phi f$ is regular, and if $f$ is regular then $phi^2 f = f$.




          Prove $phi f$ is regular via induction on the size of truncations and using the following lemma:



          Lemma 1: A sequence $f$ is regular iff for each $tinmathbbN$ the truncated sequence $f_leq t$ is regular.



          Proposition: For any sequence $f$, $phi f$ is regular.



          Proof: If $t=1$, $phi f_leq 1$ is the singleton sequence, $phi f_leq 1=(1)$, so it is trivially regular.



          Inductive step: Assume for any $s<t$, the truncated sequence $phi f_leq s$ is regular. Now show $phi f_leq t$ is regular by showing that for any truncation the defining inequality holds. In fact, it is sufficient to demonstrate the inequality for truncation up to length $t$.






          share|cite|improve this answer






















          • Yes, your definition 2 matches up to condition 2 in my second approach, see my last comment there.
            – Berci
            Aug 29 at 7:58










          Your Answer




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

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

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

          else
          createEditor();

          );

          function createEditor()
          StackExchange.prepareEditor(
          heartbeatType: 'answer',
          convertImagesToLinks: true,
          noModals: false,
          showLowRepImageUploadWarning: true,
          reputationToPostImages: 10,
          bindNavPrevention: true,
          postfix: "",
          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%2f2895510%2fshow-this-function-on-sequences-is-cyclical%23new-answer', 'question_page');

          );

          Post as a guest






























          2 Answers
          2






          active

          oldest

          votes








          2 Answers
          2






          active

          oldest

          votes









          active

          oldest

          votes






          active

          oldest

          votes








          up vote
          4
          down vote



          accepted










          Very nice observation.

          As you note, for the conjecture it suffices to prove $phi^3=phi$.



          Call sequences $a=(a_1,a_2,dots)$ and $b=(b_1,b_2,dots)$ similar, if $phi a=phi b$, i.e. for each index $n$, $ a_n$ occurs the same times among $a_1,dots,a_n $ as $ b_n$ among $b_1,dots,b_n$.

          For example, the sequences $(1,1,2,2,3,3,1), (1,1,2,2,3,3,2), (1,1,2,2,3,3,3)$ are similar.



          We have to show that $phi^2 a$ is similar to $a$ for an arbitrary sequence $a$.



          Assume it holds for $n$ long sequences, and take $a_n+1$ in $a=(a_1,dots,a_n+1)$.



          • It is either a 'new element' ($phi a(n+1)=1$), in which case the number of $1$'s that denotes the new elements, is increased by one in $phi a$, implying $phi^2 a(n+1)=|a_1,dots,a_n+1|$ which is a new element in the sequence $phi^2a$.

          • Or, it is a repeating element, and $s:=phi a(n+1)=|ile n+1:a_i=a_n+1|$.

            This $s$ occurs in $phi a$ for exactly $q$ times, where $q$ is the number of $a_i$'s occuring exactly $s$ times in $a$. (If $s$ is new, then $q=1$, and $a_n+1$ can be replaced by any of these $a_i$'s to get a similar sequence.)

            On one hand, it means $phi^2 a(n+1)=q$.

            On the other hand, $q$ already occured in $phi^2 a(1,dots,n)$ exactly $s-1=|ile n:a_i=a_n+1|$ times.


          Another approach



          Call a sequence $a$ of positive integers a regular sequence, if for all $n$,



          1. $ 1le a_nlemax_i<na_i+1$

          2. In every initial segment of $a$, the number of $x$'s is at least the number of $y$'s whenever $xle y$.

          Note that 2. can be rephrased as: if $a_n=a_i$ for some $i<n$, then $a_n$ is minimal among those sequence elements which occur exactly $(phi a)_n-1$ times in $(a_1,dots,a_n-1)$

          [because $a_n$ in $(a_1,dots,a_n)$
          has one more occurances than any of them].



          For example, $(1,2,2)$ and $(1,2,3,2)$ are not regular, while $(1,2,1)$ and $(1,2,3,1)$ are regular.



          Proposition.

          $ $ a) $ phi a$ is regular for any sequence $a$.

          $ $ b) $ $If $a$ is a regular sequence, then $phi^2 a=a$.






          share|cite|improve this answer






















          • Currently considering "level sets" of the sequence. If the original sequence $f$ has domain $D$ then for $xin D$ I let $E_x=ninmathbbNmid f(n)=x$. If you list those elements like $E_x=e_1,e_2,...$ then $f(e_n)=n$, and $phi f(e_n)$ is a constant, and furthermore $E_x$ is the only set which is mapped to this constant. Thus, $phi^2 f(e_n)=f(e_n)=n$ for all $n$.
            – M. Nestor
            Aug 26 at 23:42











          • Yes, level sets definitely play some role here around.
            – Berci
            Aug 26 at 23:46










          • Hmm.. You still have to finish the proof (of the last sentence).
            – Berci
            Aug 27 at 0:32










          • I think I can prove the case for the level sets, essentially that $(1,2,3,...)mapsto(1,1,1,...)mapsto(1,2,3,...)$, and then a larger argument piecing them together for an arbitrary sequence.
            – M. Nestor
            Aug 27 at 0:52










          • You're allowed to post your own answer. What is that argument?
            – Berci
            Aug 27 at 6:11















          up vote
          4
          down vote



          accepted










          Very nice observation.

          As you note, for the conjecture it suffices to prove $phi^3=phi$.



          Call sequences $a=(a_1,a_2,dots)$ and $b=(b_1,b_2,dots)$ similar, if $phi a=phi b$, i.e. for each index $n$, $ a_n$ occurs the same times among $a_1,dots,a_n $ as $ b_n$ among $b_1,dots,b_n$.

          For example, the sequences $(1,1,2,2,3,3,1), (1,1,2,2,3,3,2), (1,1,2,2,3,3,3)$ are similar.



          We have to show that $phi^2 a$ is similar to $a$ for an arbitrary sequence $a$.



          Assume it holds for $n$ long sequences, and take $a_n+1$ in $a=(a_1,dots,a_n+1)$.



          • It is either a 'new element' ($phi a(n+1)=1$), in which case the number of $1$'s that denotes the new elements, is increased by one in $phi a$, implying $phi^2 a(n+1)=|a_1,dots,a_n+1|$ which is a new element in the sequence $phi^2a$.

          • Or, it is a repeating element, and $s:=phi a(n+1)=|ile n+1:a_i=a_n+1|$.

            This $s$ occurs in $phi a$ for exactly $q$ times, where $q$ is the number of $a_i$'s occuring exactly $s$ times in $a$. (If $s$ is new, then $q=1$, and $a_n+1$ can be replaced by any of these $a_i$'s to get a similar sequence.)

            On one hand, it means $phi^2 a(n+1)=q$.

            On the other hand, $q$ already occured in $phi^2 a(1,dots,n)$ exactly $s-1=|ile n:a_i=a_n+1|$ times.


          Another approach



          Call a sequence $a$ of positive integers a regular sequence, if for all $n$,



          1. $ 1le a_nlemax_i<na_i+1$

          2. In every initial segment of $a$, the number of $x$'s is at least the number of $y$'s whenever $xle y$.

          Note that 2. can be rephrased as: if $a_n=a_i$ for some $i<n$, then $a_n$ is minimal among those sequence elements which occur exactly $(phi a)_n-1$ times in $(a_1,dots,a_n-1)$

          [because $a_n$ in $(a_1,dots,a_n)$
          has one more occurances than any of them].



          For example, $(1,2,2)$ and $(1,2,3,2)$ are not regular, while $(1,2,1)$ and $(1,2,3,1)$ are regular.



          Proposition.

          $ $ a) $ phi a$ is regular for any sequence $a$.

          $ $ b) $ $If $a$ is a regular sequence, then $phi^2 a=a$.






          share|cite|improve this answer






















          • Currently considering "level sets" of the sequence. If the original sequence $f$ has domain $D$ then for $xin D$ I let $E_x=ninmathbbNmid f(n)=x$. If you list those elements like $E_x=e_1,e_2,...$ then $f(e_n)=n$, and $phi f(e_n)$ is a constant, and furthermore $E_x$ is the only set which is mapped to this constant. Thus, $phi^2 f(e_n)=f(e_n)=n$ for all $n$.
            – M. Nestor
            Aug 26 at 23:42











          • Yes, level sets definitely play some role here around.
            – Berci
            Aug 26 at 23:46










          • Hmm.. You still have to finish the proof (of the last sentence).
            – Berci
            Aug 27 at 0:32










          • I think I can prove the case for the level sets, essentially that $(1,2,3,...)mapsto(1,1,1,...)mapsto(1,2,3,...)$, and then a larger argument piecing them together for an arbitrary sequence.
            – M. Nestor
            Aug 27 at 0:52










          • You're allowed to post your own answer. What is that argument?
            – Berci
            Aug 27 at 6:11













          up vote
          4
          down vote



          accepted







          up vote
          4
          down vote



          accepted






          Very nice observation.

          As you note, for the conjecture it suffices to prove $phi^3=phi$.



          Call sequences $a=(a_1,a_2,dots)$ and $b=(b_1,b_2,dots)$ similar, if $phi a=phi b$, i.e. for each index $n$, $ a_n$ occurs the same times among $a_1,dots,a_n $ as $ b_n$ among $b_1,dots,b_n$.

          For example, the sequences $(1,1,2,2,3,3,1), (1,1,2,2,3,3,2), (1,1,2,2,3,3,3)$ are similar.



          We have to show that $phi^2 a$ is similar to $a$ for an arbitrary sequence $a$.



          Assume it holds for $n$ long sequences, and take $a_n+1$ in $a=(a_1,dots,a_n+1)$.



          • It is either a 'new element' ($phi a(n+1)=1$), in which case the number of $1$'s that denotes the new elements, is increased by one in $phi a$, implying $phi^2 a(n+1)=|a_1,dots,a_n+1|$ which is a new element in the sequence $phi^2a$.

          • Or, it is a repeating element, and $s:=phi a(n+1)=|ile n+1:a_i=a_n+1|$.

            This $s$ occurs in $phi a$ for exactly $q$ times, where $q$ is the number of $a_i$'s occuring exactly $s$ times in $a$. (If $s$ is new, then $q=1$, and $a_n+1$ can be replaced by any of these $a_i$'s to get a similar sequence.)

            On one hand, it means $phi^2 a(n+1)=q$.

            On the other hand, $q$ already occured in $phi^2 a(1,dots,n)$ exactly $s-1=|ile n:a_i=a_n+1|$ times.


          Another approach



          Call a sequence $a$ of positive integers a regular sequence, if for all $n$,



          1. $ 1le a_nlemax_i<na_i+1$

          2. In every initial segment of $a$, the number of $x$'s is at least the number of $y$'s whenever $xle y$.

          Note that 2. can be rephrased as: if $a_n=a_i$ for some $i<n$, then $a_n$ is minimal among those sequence elements which occur exactly $(phi a)_n-1$ times in $(a_1,dots,a_n-1)$

          [because $a_n$ in $(a_1,dots,a_n)$
          has one more occurances than any of them].



          For example, $(1,2,2)$ and $(1,2,3,2)$ are not regular, while $(1,2,1)$ and $(1,2,3,1)$ are regular.



          Proposition.

          $ $ a) $ phi a$ is regular for any sequence $a$.

          $ $ b) $ $If $a$ is a regular sequence, then $phi^2 a=a$.






          share|cite|improve this answer














          Very nice observation.

          As you note, for the conjecture it suffices to prove $phi^3=phi$.



          Call sequences $a=(a_1,a_2,dots)$ and $b=(b_1,b_2,dots)$ similar, if $phi a=phi b$, i.e. for each index $n$, $ a_n$ occurs the same times among $a_1,dots,a_n $ as $ b_n$ among $b_1,dots,b_n$.

          For example, the sequences $(1,1,2,2,3,3,1), (1,1,2,2,3,3,2), (1,1,2,2,3,3,3)$ are similar.



          We have to show that $phi^2 a$ is similar to $a$ for an arbitrary sequence $a$.



          Assume it holds for $n$ long sequences, and take $a_n+1$ in $a=(a_1,dots,a_n+1)$.



          • It is either a 'new element' ($phi a(n+1)=1$), in which case the number of $1$'s that denotes the new elements, is increased by one in $phi a$, implying $phi^2 a(n+1)=|a_1,dots,a_n+1|$ which is a new element in the sequence $phi^2a$.

          • Or, it is a repeating element, and $s:=phi a(n+1)=|ile n+1:a_i=a_n+1|$.

            This $s$ occurs in $phi a$ for exactly $q$ times, where $q$ is the number of $a_i$'s occuring exactly $s$ times in $a$. (If $s$ is new, then $q=1$, and $a_n+1$ can be replaced by any of these $a_i$'s to get a similar sequence.)

            On one hand, it means $phi^2 a(n+1)=q$.

            On the other hand, $q$ already occured in $phi^2 a(1,dots,n)$ exactly $s-1=|ile n:a_i=a_n+1|$ times.


          Another approach



          Call a sequence $a$ of positive integers a regular sequence, if for all $n$,



          1. $ 1le a_nlemax_i<na_i+1$

          2. In every initial segment of $a$, the number of $x$'s is at least the number of $y$'s whenever $xle y$.

          Note that 2. can be rephrased as: if $a_n=a_i$ for some $i<n$, then $a_n$ is minimal among those sequence elements which occur exactly $(phi a)_n-1$ times in $(a_1,dots,a_n-1)$

          [because $a_n$ in $(a_1,dots,a_n)$
          has one more occurances than any of them].



          For example, $(1,2,2)$ and $(1,2,3,2)$ are not regular, while $(1,2,1)$ and $(1,2,3,1)$ are regular.



          Proposition.

          $ $ a) $ phi a$ is regular for any sequence $a$.

          $ $ b) $ $If $a$ is a regular sequence, then $phi^2 a=a$.







          share|cite|improve this answer














          share|cite|improve this answer



          share|cite|improve this answer








          edited Aug 31 at 22:22

























          answered Aug 26 at 22:54









          Berci

          57k23570




          57k23570











          • Currently considering "level sets" of the sequence. If the original sequence $f$ has domain $D$ then for $xin D$ I let $E_x=ninmathbbNmid f(n)=x$. If you list those elements like $E_x=e_1,e_2,...$ then $f(e_n)=n$, and $phi f(e_n)$ is a constant, and furthermore $E_x$ is the only set which is mapped to this constant. Thus, $phi^2 f(e_n)=f(e_n)=n$ for all $n$.
            – M. Nestor
            Aug 26 at 23:42











          • Yes, level sets definitely play some role here around.
            – Berci
            Aug 26 at 23:46










          • Hmm.. You still have to finish the proof (of the last sentence).
            – Berci
            Aug 27 at 0:32










          • I think I can prove the case for the level sets, essentially that $(1,2,3,...)mapsto(1,1,1,...)mapsto(1,2,3,...)$, and then a larger argument piecing them together for an arbitrary sequence.
            – M. Nestor
            Aug 27 at 0:52










          • You're allowed to post your own answer. What is that argument?
            – Berci
            Aug 27 at 6:11

















          • Currently considering "level sets" of the sequence. If the original sequence $f$ has domain $D$ then for $xin D$ I let $E_x=ninmathbbNmid f(n)=x$. If you list those elements like $E_x=e_1,e_2,...$ then $f(e_n)=n$, and $phi f(e_n)$ is a constant, and furthermore $E_x$ is the only set which is mapped to this constant. Thus, $phi^2 f(e_n)=f(e_n)=n$ for all $n$.
            – M. Nestor
            Aug 26 at 23:42











          • Yes, level sets definitely play some role here around.
            – Berci
            Aug 26 at 23:46










          • Hmm.. You still have to finish the proof (of the last sentence).
            – Berci
            Aug 27 at 0:32










          • I think I can prove the case for the level sets, essentially that $(1,2,3,...)mapsto(1,1,1,...)mapsto(1,2,3,...)$, and then a larger argument piecing them together for an arbitrary sequence.
            – M. Nestor
            Aug 27 at 0:52










          • You're allowed to post your own answer. What is that argument?
            – Berci
            Aug 27 at 6:11
















          Currently considering "level sets" of the sequence. If the original sequence $f$ has domain $D$ then for $xin D$ I let $E_x=ninmathbbNmid f(n)=x$. If you list those elements like $E_x=e_1,e_2,...$ then $f(e_n)=n$, and $phi f(e_n)$ is a constant, and furthermore $E_x$ is the only set which is mapped to this constant. Thus, $phi^2 f(e_n)=f(e_n)=n$ for all $n$.
          – M. Nestor
          Aug 26 at 23:42





          Currently considering "level sets" of the sequence. If the original sequence $f$ has domain $D$ then for $xin D$ I let $E_x=ninmathbbNmid f(n)=x$. If you list those elements like $E_x=e_1,e_2,...$ then $f(e_n)=n$, and $phi f(e_n)$ is a constant, and furthermore $E_x$ is the only set which is mapped to this constant. Thus, $phi^2 f(e_n)=f(e_n)=n$ for all $n$.
          – M. Nestor
          Aug 26 at 23:42













          Yes, level sets definitely play some role here around.
          – Berci
          Aug 26 at 23:46




          Yes, level sets definitely play some role here around.
          – Berci
          Aug 26 at 23:46












          Hmm.. You still have to finish the proof (of the last sentence).
          – Berci
          Aug 27 at 0:32




          Hmm.. You still have to finish the proof (of the last sentence).
          – Berci
          Aug 27 at 0:32












          I think I can prove the case for the level sets, essentially that $(1,2,3,...)mapsto(1,1,1,...)mapsto(1,2,3,...)$, and then a larger argument piecing them together for an arbitrary sequence.
          – M. Nestor
          Aug 27 at 0:52




          I think I can prove the case for the level sets, essentially that $(1,2,3,...)mapsto(1,1,1,...)mapsto(1,2,3,...)$, and then a larger argument piecing them together for an arbitrary sequence.
          – M. Nestor
          Aug 27 at 0:52












          You're allowed to post your own answer. What is that argument?
          – Berci
          Aug 27 at 6:11





          You're allowed to post your own answer. What is that argument?
          – Berci
          Aug 27 at 6:11











          up vote
          1
          down vote













          Solution in progress



          For a sequence $f$ define



          1. the pre-image of $f$ denoted $f^-1(a):=ninmathbbNmid f(n)=a$.


          2. a truncation of $f$ at $t$, denoted $f_leq t$, is the sequence with the domain restricted to $ninmathbbNmid nleq t$.


          Then by $f_leq t^-1(n)$ denote $kinmathbbNmid kleq t,f(k)=f(n)$. Here is a new definition for $phi$:




          Definition 1: For any sequence $f$, define $phi f(n) := textcard f_leq n^-1(n) $




          The following definition and conjecture describes the sequences that $phi$ will produce:




          Definition 2: A sequence $f$ with range $mathbbN$ is regular if for any truncation at $tinmathbbN$, $$textcard f_leq
          t^-1(1)geqtextcard f_leq t^-1(2)geqtextcard f_leq
          t^-1(3)geq...$$




          That is, for any truncation of the sequence $f$, the number of $1$s is $geq$ the number of $2$s, which is $geq$ the number of $3$s and so on, which ensures the sequence is counting properly. I am nearly certain this definition is equivalent to those in comments.




          Conjecture: If $f$ is any sequence then $phi f$ is regular, and if $f$ is regular then $phi^2 f = f$.




          Prove $phi f$ is regular via induction on the size of truncations and using the following lemma:



          Lemma 1: A sequence $f$ is regular iff for each $tinmathbbN$ the truncated sequence $f_leq t$ is regular.



          Proposition: For any sequence $f$, $phi f$ is regular.



          Proof: If $t=1$, $phi f_leq 1$ is the singleton sequence, $phi f_leq 1=(1)$, so it is trivially regular.



          Inductive step: Assume for any $s<t$, the truncated sequence $phi f_leq s$ is regular. Now show $phi f_leq t$ is regular by showing that for any truncation the defining inequality holds. In fact, it is sufficient to demonstrate the inequality for truncation up to length $t$.






          share|cite|improve this answer






















          • Yes, your definition 2 matches up to condition 2 in my second approach, see my last comment there.
            – Berci
            Aug 29 at 7:58














          up vote
          1
          down vote













          Solution in progress



          For a sequence $f$ define



          1. the pre-image of $f$ denoted $f^-1(a):=ninmathbbNmid f(n)=a$.


          2. a truncation of $f$ at $t$, denoted $f_leq t$, is the sequence with the domain restricted to $ninmathbbNmid nleq t$.


          Then by $f_leq t^-1(n)$ denote $kinmathbbNmid kleq t,f(k)=f(n)$. Here is a new definition for $phi$:




          Definition 1: For any sequence $f$, define $phi f(n) := textcard f_leq n^-1(n) $




          The following definition and conjecture describes the sequences that $phi$ will produce:




          Definition 2: A sequence $f$ with range $mathbbN$ is regular if for any truncation at $tinmathbbN$, $$textcard f_leq
          t^-1(1)geqtextcard f_leq t^-1(2)geqtextcard f_leq
          t^-1(3)geq...$$




          That is, for any truncation of the sequence $f$, the number of $1$s is $geq$ the number of $2$s, which is $geq$ the number of $3$s and so on, which ensures the sequence is counting properly. I am nearly certain this definition is equivalent to those in comments.




          Conjecture: If $f$ is any sequence then $phi f$ is regular, and if $f$ is regular then $phi^2 f = f$.




          Prove $phi f$ is regular via induction on the size of truncations and using the following lemma:



          Lemma 1: A sequence $f$ is regular iff for each $tinmathbbN$ the truncated sequence $f_leq t$ is regular.



          Proposition: For any sequence $f$, $phi f$ is regular.



          Proof: If $t=1$, $phi f_leq 1$ is the singleton sequence, $phi f_leq 1=(1)$, so it is trivially regular.



          Inductive step: Assume for any $s<t$, the truncated sequence $phi f_leq s$ is regular. Now show $phi f_leq t$ is regular by showing that for any truncation the defining inequality holds. In fact, it is sufficient to demonstrate the inequality for truncation up to length $t$.






          share|cite|improve this answer






















          • Yes, your definition 2 matches up to condition 2 in my second approach, see my last comment there.
            – Berci
            Aug 29 at 7:58












          up vote
          1
          down vote










          up vote
          1
          down vote









          Solution in progress



          For a sequence $f$ define



          1. the pre-image of $f$ denoted $f^-1(a):=ninmathbbNmid f(n)=a$.


          2. a truncation of $f$ at $t$, denoted $f_leq t$, is the sequence with the domain restricted to $ninmathbbNmid nleq t$.


          Then by $f_leq t^-1(n)$ denote $kinmathbbNmid kleq t,f(k)=f(n)$. Here is a new definition for $phi$:




          Definition 1: For any sequence $f$, define $phi f(n) := textcard f_leq n^-1(n) $




          The following definition and conjecture describes the sequences that $phi$ will produce:




          Definition 2: A sequence $f$ with range $mathbbN$ is regular if for any truncation at $tinmathbbN$, $$textcard f_leq
          t^-1(1)geqtextcard f_leq t^-1(2)geqtextcard f_leq
          t^-1(3)geq...$$




          That is, for any truncation of the sequence $f$, the number of $1$s is $geq$ the number of $2$s, which is $geq$ the number of $3$s and so on, which ensures the sequence is counting properly. I am nearly certain this definition is equivalent to those in comments.




          Conjecture: If $f$ is any sequence then $phi f$ is regular, and if $f$ is regular then $phi^2 f = f$.




          Prove $phi f$ is regular via induction on the size of truncations and using the following lemma:



          Lemma 1: A sequence $f$ is regular iff for each $tinmathbbN$ the truncated sequence $f_leq t$ is regular.



          Proposition: For any sequence $f$, $phi f$ is regular.



          Proof: If $t=1$, $phi f_leq 1$ is the singleton sequence, $phi f_leq 1=(1)$, so it is trivially regular.



          Inductive step: Assume for any $s<t$, the truncated sequence $phi f_leq s$ is regular. Now show $phi f_leq t$ is regular by showing that for any truncation the defining inequality holds. In fact, it is sufficient to demonstrate the inequality for truncation up to length $t$.






          share|cite|improve this answer














          Solution in progress



          For a sequence $f$ define



          1. the pre-image of $f$ denoted $f^-1(a):=ninmathbbNmid f(n)=a$.


          2. a truncation of $f$ at $t$, denoted $f_leq t$, is the sequence with the domain restricted to $ninmathbbNmid nleq t$.


          Then by $f_leq t^-1(n)$ denote $kinmathbbNmid kleq t,f(k)=f(n)$. Here is a new definition for $phi$:




          Definition 1: For any sequence $f$, define $phi f(n) := textcard f_leq n^-1(n) $




          The following definition and conjecture describes the sequences that $phi$ will produce:




          Definition 2: A sequence $f$ with range $mathbbN$ is regular if for any truncation at $tinmathbbN$, $$textcard f_leq
          t^-1(1)geqtextcard f_leq t^-1(2)geqtextcard f_leq
          t^-1(3)geq...$$




          That is, for any truncation of the sequence $f$, the number of $1$s is $geq$ the number of $2$s, which is $geq$ the number of $3$s and so on, which ensures the sequence is counting properly. I am nearly certain this definition is equivalent to those in comments.




          Conjecture: If $f$ is any sequence then $phi f$ is regular, and if $f$ is regular then $phi^2 f = f$.




          Prove $phi f$ is regular via induction on the size of truncations and using the following lemma:



          Lemma 1: A sequence $f$ is regular iff for each $tinmathbbN$ the truncated sequence $f_leq t$ is regular.



          Proposition: For any sequence $f$, $phi f$ is regular.



          Proof: If $t=1$, $phi f_leq 1$ is the singleton sequence, $phi f_leq 1=(1)$, so it is trivially regular.



          Inductive step: Assume for any $s<t$, the truncated sequence $phi f_leq s$ is regular. Now show $phi f_leq t$ is regular by showing that for any truncation the defining inequality holds. In fact, it is sufficient to demonstrate the inequality for truncation up to length $t$.







          share|cite|improve this answer














          share|cite|improve this answer



          share|cite|improve this answer








          edited Aug 28 at 18:16

























          answered Aug 28 at 18:01









          M. Nestor

          4519




          4519











          • Yes, your definition 2 matches up to condition 2 in my second approach, see my last comment there.
            – Berci
            Aug 29 at 7:58
















          • Yes, your definition 2 matches up to condition 2 in my second approach, see my last comment there.
            – Berci
            Aug 29 at 7:58















          Yes, your definition 2 matches up to condition 2 in my second approach, see my last comment there.
          – Berci
          Aug 29 at 7:58




          Yes, your definition 2 matches up to condition 2 in my second approach, see my last comment there.
          – Berci
          Aug 29 at 7:58

















           

          draft saved


          draft discarded















































           


          draft saved


          draft discarded














          StackExchange.ready(
          function ()
          StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2895510%2fshow-this-function-on-sequences-is-cyclical%23new-answer', 'question_page');

          );

          Post as a guest













































































          這個網誌中的熱門文章

          How to combine Bézier curves to a surface?

          Why am i infinitely getting the same tweet with the Twitter Search API?

          Carbon dioxide