Skip to main content
deleted 1140 characters in body; edited title; edited title
Source Link
Xuemei
  • 141
  • 5

Any proofs for perfect matchings with assigned encoding color number are different after removing its How to prove polynomial inequality encoded from 1-factorizations?factors in $K_{2n}$

some Definition:

For a Graph $G=(V,E)$ (here $V=\{v_{m}|m=1,2,...,n\}, E=\{e_{ij}|i,j=1,2,...,n\}$), we define one monochrome color stands for one disjoint perfect matchings and each monochrome color is represented as an integer number $y$. Thus every monochrome colored edge $e_{ij}$ can be seen as lablled by $yy$, which also means that the number list for vertices is $yy$.

One example: let us assume red means number 0. For a perfect matching of $K_{6}$ in red Let ($e_{1,2}\ ,\ e_{3,4}\ ,\ e_{5,6}$)$G=(V,E)$ be a complete graph , the number list$K_{2n}$ and it has $X_{i}$ for vertices$m$ 1-factors ($v_{1}$,$v_{2}$,$v_{3}$,$v_{4}$,$v_{5}$$f_{i,(i=1,\dots,m)}$,$v_{6}$) is where $X_{i}$ =($v_{1}$,$v_{2}$,$v_{3}$,$v_{4}$,$v_{5}$,$v_{6}$)=(0,0,0,0,0,0)$m=\frac{(2n)!}{n!2^n}$.

See the below figure for more understanding:

enter image description hereSome definition:

other Information:

  • The edge set of$F=\{f_{1},f_{2},...,f_{2n-1}\}$ is one $K_{2n}$ can be divided into1-factorization in ($2n-1$) disjoint perfect matching, which$K_{2n}$.
  • $S(F)$ is the max number of disjoint perfect matchings1-factors in one (or$F$; and $max \ disjoint\ PM(𝐾_{2n})=2n-1$)$S(F)=2n-1$.
  • The$N(F)$ is the number of distinct 1-factorizations of a complete graphin $K_{2n}$ (we call it as $\#$ways for;
  • If $K_{2n}$) can been obtained from$N(F)=k$, we use A000438$F_{1}$,$F_{2}$,. (example: $\#$ways for..,$F_{k}$ to indicate distinct $K_{6}$ is$F$ $6$).
  • For one certain color setting (or 1-factorizations), all perfect matchings of $K_{2n}$ can be written as term $T_{apm}=T_{d}+T_{r}$, wherewhen $T_{d}=\sum_{i=1}^{2n-1}X_{i}$$n=1,2,3,4$, we know $T_{r}=\sum_{i=2n}^{PM(K_{2n})}X_{i}$$N(F)=1,1,6,6240$; see A000438.

Question: For each $f_{i}$ in $K_{2n}$, we label the edges of $f_{i}$ with numbers 1,2,...,(2n-1). For each $f_{i}$ in $F$, none of them are labeled with the same number (which means disjoint 1-factors).

  1. Let us take the complete graph $K_{6}$ as an example, we know #$PM(K_{6})=15$, $max \ disjoint\ PM(𝐾_{6})=5$ and $\#ways$ $K_{6}=6$. We list one way of color settings for 5 disjoint perfect matchings as the following figure:

enter image description here Now we associate with vertex $v$ in $K_{2n}$ with a word (not number) $r$ such that $r$ is the label of the edge incident with $v$ in $f_{i}$ of $K_{2n}$. It's straightforward to check that the resulting code {$r,v\in V$} has desired parameter from $1,2,...,(2n-1)$. We label the coded word in one $f_{i}$ as $W(f_{i})=r_{1}r_{2}...r_{2n}$ (corresponding to the sequence arrangement as $v_{1}v_{2}...v_{2n}$).

  • Let us remove the term $T_{d}=\sum_{i=1}^{5}X_{i}$, which comes from the 5 disjoint perfect macthings. And we define the rest term as $T_{r}=\sum_{i=6}^{15}X_{i}$.

notes: example $T_{r_{1}}=\sum_{i=1}^{3}X_{i}$ issee the form offollowing two examples for description:

$f_{x}=\{(v_{1}v_{4}),(v_{2}v_{6}),(v_{3}v_{5})\}=\{(11),(22),(33)\}$ $(0,0,0,0,0,0)+(1,1,1,1,1,1)+(2,2,2,2,2,2)$ and should not equal to adding all elements that gives: $(3,3,3,3,3,3)$. Thus we mean$W(f_{x})=r_{1}r_{2}...r_{6}=123132$ $f_{x}=\{(v_{1}v_{6}),(v_{2}v_{5}),(v_{3}v_{4})\}=\{(22),(44),(55)\}$ $(0,0,0,0,0,0)+(1,1,1,1,1,1)+(2,2,2,2,2,2)$: !=$(3,3,3,3,3,3)$$W(f_{x})=r_{1}r_{2}...r_{6}=245542$


We separate $m$ 1-factors $f_{i}$ into two groups:

  • There are 6 ways of color setting for the 5 disjoint perfect matchings, thus there will be sixgroup1: $T_{r_{j}}${$f_{i|i=1,2,...,2n-1}$} which forms $F$ ($j=6$disjoint 1-factors in $K_{2n}$).
  • And then we do the permutation of number list $L=\{0,1,2,3,4\}$. The monochrome color listgroup2: {red, blue, green, yellow, gray$f_{i|i=2n,2n+1,...,m}$} encodes in one permutation. The total number iswhich forms Length[Permutations[L]]=120.
  • There will be 120*6=720$F'$ $T_{r_{k}}, k=1,...,720$, and we find all(remaining 1-factors in $T_{r_{k}}$ is different.$K_{2n}$)

They form two polynomials: $Y(F)=\sum_{i=1}^{(2n-1)} W(f_{i})$ and $Y(F')=\sum_{i=2n}^m W(f_{i})$. Thus there will be $N(F)$ cases for $Y(F')$, which we can use $Y(F')_{1}$, $Y(F')_{2}$, ..., $Y(F')_{N(F)}$to indicate.

So why all the $T_{r_{k}}, k=720$ are different? Is there any proof or theorem instead to write a program to check all the cases?Question:

  • For $K_{8}$, we know #$PM(K_{8})=105$, $max \ disjoint\ PM(𝐾_{8})=7$, $\#ways$ $K_{8}=6240$. Then for one way of 7 monochrome color setting, there will be $T_{d}=\sum_{i=1}^{7}X_{i}$ and $T_{r}=\sum_{i=8}^{105}X_{i}$.

  • Again we do the permutation of number list $L=\{0,1,2,3,4,5,6\}$, so there will be Length[Permutations[L]]=5040 possibilities.

  • There will be 5040*6240=31449600 cases $T_{r_{k}}, k=1,...,31449600$. Will all $T_{r_{k}}$ also be different?

Thank you very muchFor 1-factorization $F$, one can permute the numbers from ($1,2,...,2n-1$) to label each $f_{i}$ in advance! If$F$, thus there are something unclearwill be $(2n-1)!$ label settings. In addition, please let methere are in total $N(F)$ 1-factorizations in $K_{2n}$. Therefore we know there will be $t=N(F)*(2n-1)!$ cases for $Y(F')$. Thank you!

$\exists i,j\in [1,t]$ and $i\not\equiv j$, $Y(F')_{i}==Y(F')_{j}$?

Any proofs for perfect matchings with assigned encoding color number are different after removing its 1-factorizations?

some Definition:

For a Graph $G=(V,E)$ (here $V=\{v_{m}|m=1,2,...,n\}, E=\{e_{ij}|i,j=1,2,...,n\}$), we define one monochrome color stands for one disjoint perfect matchings and each monochrome color is represented as an integer number $y$. Thus every monochrome colored edge $e_{ij}$ can be seen as lablled by $yy$, which also means that the number list for vertices is $yy$.

One example: let us assume red means number 0. For a perfect matching of $K_{6}$ in red ($e_{1,2}\ ,\ e_{3,4}\ ,\ e_{5,6}$) , the number list $X_{i}$ for vertices ($v_{1}$,$v_{2}$,$v_{3}$,$v_{4}$,$v_{5}$,$v_{6}$) is $X_{i}$ =($v_{1}$,$v_{2}$,$v_{3}$,$v_{4}$,$v_{5}$,$v_{6}$)=(0,0,0,0,0,0).

See the below figure for more understanding:

enter image description here

other Information:

  • The edge set of $K_{2n}$ can be divided into ($2n-1$) disjoint perfect matching, which is the max number of disjoint perfect matchings (or $max \ disjoint\ PM(𝐾_{2n})=2n-1$).
  • The number of 1-factorizations of a complete graph $K_{2n}$ (we call it as $\#$ways for $K_{2n}$) can been obtained from A000438. (example: $\#$ways for $K_{6}$ is $6$)
  • For one certain color setting (or 1-factorizations), all perfect matchings of $K_{2n}$ can be written as term $T_{apm}=T_{d}+T_{r}$, where $T_{d}=\sum_{i=1}^{2n-1}X_{i}$, $T_{r}=\sum_{i=2n}^{PM(K_{2n})}X_{i}$.

Question:

  1. Let us take the complete graph $K_{6}$ as an example, we know #$PM(K_{6})=15$, $max \ disjoint\ PM(𝐾_{6})=5$ and $\#ways$ $K_{6}=6$. We list one way of color settings for 5 disjoint perfect matchings as the following figure:

enter image description here

  • Let us remove the term $T_{d}=\sum_{i=1}^{5}X_{i}$, which comes from the 5 disjoint perfect macthings. And we define the rest term as $T_{r}=\sum_{i=6}^{15}X_{i}$.

notes: example $T_{r_{1}}=\sum_{i=1}^{3}X_{i}$ is the form of $(0,0,0,0,0,0)+(1,1,1,1,1,1)+(2,2,2,2,2,2)$ and should not equal to adding all elements that gives $(3,3,3,3,3,3)$. Thus we mean $(0,0,0,0,0,0)+(1,1,1,1,1,1)+(2,2,2,2,2,2)$ !=$(3,3,3,3,3,3)$

  • There are 6 ways of color setting for the 5 disjoint perfect matchings, thus there will be six $T_{r_{j}}$ ($j=6$).
  • And then we do the permutation of number list $L=\{0,1,2,3,4\}$. The monochrome color list {red, blue, green, yellow, gray} encodes in one permutation. The total number is Length[Permutations[L]]=120.
  • There will be 120*6=720 $T_{r_{k}}, k=1,...,720$, and we find all $T_{r_{k}}$ is different.

So why all the $T_{r_{k}}, k=720$ are different? Is there any proof or theorem instead to write a program to check all the cases?

  • For $K_{8}$, we know #$PM(K_{8})=105$, $max \ disjoint\ PM(𝐾_{8})=7$, $\#ways$ $K_{8}=6240$. Then for one way of 7 monochrome color setting, there will be $T_{d}=\sum_{i=1}^{7}X_{i}$ and $T_{r}=\sum_{i=8}^{105}X_{i}$.

  • Again we do the permutation of number list $L=\{0,1,2,3,4,5,6\}$, so there will be Length[Permutations[L]]=5040 possibilities.

  • There will be 5040*6240=31449600 cases $T_{r_{k}}, k=1,...,31449600$. Will all $T_{r_{k}}$ also be different?

Thank you very much in advance! If there are something unclear, please let me know. Thank you!

How to prove polynomial inequality encoded from 1-factors in $K_{2n}$

Let $G=(V,E)$ be a complete graph $K_{2n}$ and it has $m$ 1-factors $f_{i,(i=1,\dots,m)}$, where $m=\frac{(2n)!}{n!2^n}$.

Some definition:

  • $F=\{f_{1},f_{2},...,f_{2n-1}\}$ is one 1-factorization in $K_{2n}$.
  • $S(F)$ is the number of 1-factors in one $F$; and $S(F)=2n-1$.
  • $N(F)$ is the number of distinct 1-factorizations in $K_{2n}$;
  • If $N(F)=k$, we use $F_{1}$,$F_{2}$,...,$F_{k}$ to indicate distinct $F$ .
  • For $K_{2n}$, when $n=1,2,3,4$, we know $N(F)=1,1,6,6240$; see A000438.

For each $f_{i}$ in $K_{2n}$, we label the edges of $f_{i}$ with numbers 1,2,...,(2n-1). For each $f_{i}$ in $F$, none of them are labeled with the same number (which means disjoint 1-factors).

Now we associate with vertex $v$ in $K_{2n}$ with a word (not number) $r$ such that $r$ is the label of the edge incident with $v$ in $f_{i}$ of $K_{2n}$. It's straightforward to check that the resulting code {$r,v\in V$} has desired parameter from $1,2,...,(2n-1)$. We label the coded word in one $f_{i}$ as $W(f_{i})=r_{1}r_{2}...r_{2n}$ (corresponding to the sequence arrangement as $v_{1}v_{2}...v_{2n}$).

see the following two examples for description:

$f_{x}=\{(v_{1}v_{4}),(v_{2}v_{6}),(v_{3}v_{5})\}=\{(11),(22),(33)\}$ : $W(f_{x})=r_{1}r_{2}...r_{6}=123132$ $f_{x}=\{(v_{1}v_{6}),(v_{2}v_{5}),(v_{3}v_{4})\}=\{(22),(44),(55)\}$ : $W(f_{x})=r_{1}r_{2}...r_{6}=245542$


We separate $m$ 1-factors $f_{i}$ into two groups:

  • group1: {$f_{i|i=1,2,...,2n-1}$} which forms $F$ (disjoint 1-factors in $K_{2n}$)
  • group2: {$f_{i|i=2n,2n+1,...,m}$} which forms $F'$ (remaining 1-factors in $K_{2n}$)

They form two polynomials: $Y(F)=\sum_{i=1}^{(2n-1)} W(f_{i})$ and $Y(F')=\sum_{i=2n}^m W(f_{i})$. Thus there will be $N(F)$ cases for $Y(F')$, which we can use $Y(F')_{1}$, $Y(F')_{2}$, ..., $Y(F')_{N(F)}$to indicate.

Question:

For 1-factorization $F$, one can permute the numbers from ($1,2,...,2n-1$) to label each $f_{i}$ in $F$, thus there will be $(2n-1)!$ label settings. In addition, there are in total $N(F)$ 1-factorizations in $K_{2n}$. Therefore we know there will be $t=N(F)*(2n-1)!$ cases for $Y(F')$.

$\exists i,j\in [1,t]$ and $i\not\equiv j$, $Y(F')_{i}==Y(F')_{j}$?

edited title
Link
Xuemei
  • 141
  • 5

Any proofs for perfect matchings with assigned encoding color number isare different after removing its 1-factorizations?

Source Link
Xuemei
  • 141
  • 5

Any proofs for perfect matchings with assigned encoding color number is different after removing its 1-factorizations?

some Definition:

For a Graph $G=(V,E)$ (here $V=\{v_{m}|m=1,2,...,n\}, E=\{e_{ij}|i,j=1,2,...,n\}$), we define one monochrome color stands for one disjoint perfect matchings and each monochrome color is represented as an integer number $y$. Thus every monochrome colored edge $e_{ij}$ can be seen as lablled by $yy$, which also means that the number list for vertices is $yy$.

One example: let us assume red means number 0. For a perfect matching of $K_{6}$ in red ($e_{1,2}\ ,\ e_{3,4}\ ,\ e_{5,6}$) , the number list $X_{i}$ for vertices ($v_{1}$,$v_{2}$,$v_{3}$,$v_{4}$,$v_{5}$,$v_{6}$) is $X_{i}$ =($v_{1}$,$v_{2}$,$v_{3}$,$v_{4}$,$v_{5}$,$v_{6}$)=(0,0,0,0,0,0).

See the below figure for more understanding:

enter image description here

other Information:

  • The edge set of $K_{2n}$ can be divided into ($2n-1$) disjoint perfect matching, which is the max number of disjoint perfect matchings (or $max \ disjoint\ PM(𝐾_{2n})=2n-1$).
  • The number of 1-factorizations of a complete graph $K_{2n}$ (we call it as $\#$ways for $K_{2n}$) can been obtained from A000438. (example: $\#$ways for $K_{6}$ is $6$)
  • For one certain color setting (or 1-factorizations), all perfect matchings of $K_{2n}$ can be written as term $T_{apm}=T_{d}+T_{r}$, where $T_{d}=\sum_{i=1}^{2n-1}X_{i}$, $T_{r}=\sum_{i=2n}^{PM(K_{2n})}X_{i}$.

Question:

  1. Let us take the complete graph $K_{6}$ as an example, we know #$PM(K_{6})=15$, $max \ disjoint\ PM(𝐾_{6})=5$ and $\#ways$ $K_{6}=6$. We list one way of color settings for 5 disjoint perfect matchings as the following figure:

enter image description here

  • Let us remove the term $T_{d}=\sum_{i=1}^{5}X_{i}$, which comes from the 5 disjoint perfect macthings. And we define the rest term as $T_{r}=\sum_{i=6}^{15}X_{i}$.

notes: example $T_{r_{1}}=\sum_{i=1}^{3}X_{i}$ is the form of $(0,0,0,0,0,0)+(1,1,1,1,1,1)+(2,2,2,2,2,2)$ and should not equal to adding all elements that gives $(3,3,3,3,3,3)$. Thus we mean $(0,0,0,0,0,0)+(1,1,1,1,1,1)+(2,2,2,2,2,2)$ !=$(3,3,3,3,3,3)$

  • There are 6 ways of color setting for the 5 disjoint perfect matchings, thus there will be six $T_{r_{j}}$ ($j=6$).
  • And then we do the permutation of number list $L=\{0,1,2,3,4\}$. The monochrome color list {red, blue, green, yellow, gray} encodes in one permutation. The total number is Length[Permutations[L]]=120.
  • There will be 120*6=720 $T_{r_{k}}, k=1,...,720$, and we find all $T_{r_{k}}$ is different.

So why all the $T_{r_{k}}, k=720$ are different? Is there any proof or theorem instead to write a program to check all the cases?

  • For $K_{8}$, we know #$PM(K_{8})=105$, $max \ disjoint\ PM(𝐾_{8})=7$, $\#ways$ $K_{8}=6240$. Then for one way of 7 monochrome color setting, there will be $T_{d}=\sum_{i=1}^{7}X_{i}$ and $T_{r}=\sum_{i=8}^{105}X_{i}$.

  • Again we do the permutation of number list $L=\{0,1,2,3,4,5,6\}$, so there will be Length[Permutations[L]]=5040 possibilities.

  • There will be 5040*6240=31449600 cases $T_{r_{k}}, k=1,...,31449600$. Will all $T_{r_{k}}$ also be different?

Thank you very much in advance! If there are something unclear, please let me know. Thank you!