Difference between revisions of "Young tableau"
(→Comments: image) |
m (tex done) |
||
(3 intermediate revisions by one other user not shown) | |||
Line 1: | Line 1: | ||
+ | |||
''of order $m$'' | ''of order $m$'' | ||
− | A [[Young diagram]] of order $m$ in whose cells the different numbers $1,\ldots,m$ have been inserted in some order, e.g. | + | A [[Young diagram]] of order $m$ in whose cells the different numbers $1,\ldots,m$ have been inserted in some order, ''e.g.'' |
− | + | ||
− | + | <pre style="font-family: monospace;color:black"> | |
− | + | ┌───┬───┬───┬───┐ | |
+ | │ 5 │ 7 │ 9 │ 4 │ | ||
+ | ├───┼───┼───┼───┘ | ||
+ | │ 8 │ 2 │ 1 │ | ||
+ | ├───┼───┴───┘ | ||
+ | │ 3 │ | ||
+ | ├───┤ | ||
+ | │ 6 │ | ||
+ | └───┘ | ||
+ | </pre> | ||
− | A Young tableau is called standard if in each row and in each column the numbers occur in increasing order. The number of all Young | + | A Young tableau is called standard if in each row and in each column the numbers occur in increasing order. The number of all Young tableaux for a given Young diagram $t$ of order $m$ is equal to $m!$ and the number of standard Young tableaux is |
$$ | $$ | ||
\frac{m!}{\prod\lambda_{ij}} | \frac{m!}{\prod\lambda_{ij}} | ||
Line 16: | Line 26: | ||
In Western literature the phrase Ferrers diagram is also used for a Young diagram. In the Russian literature the phrase "Young tableau" ( "Yunga tablitsa" ) and "Young diagram" ( "Yunga diagramma" ) are used precisely in the opposite way, with "tablitsa" referring to the pictorial representation of a partition and "diagramma" being a filled-in "tablitsa" . | In Western literature the phrase Ferrers diagram is also used for a Young diagram. In the Russian literature the phrase "Young tableau" ( "Yunga tablitsa" ) and "Young diagram" ( "Yunga diagramma" ) are used precisely in the opposite way, with "tablitsa" referring to the pictorial representation of a partition and "diagramma" being a filled-in "tablitsa" . | ||
− | Let $\kappa$ denote a [[partition]] of $m$ ( | + | Let $\kappa$ denote a [[partition]] of $m$ ($\kappa=(\kappa_1,\ldots,\kappa_t)$, $\kappa_i \in \{0,1,\ldots\}$, $\kappa_1+\cdots+\kappa_t=m$) as well as its corresponding [[Young diagram]], its pictorial representation. Let $\lambda$ be a second partition of $m$. A $\kappa$-tableau of type $\lambda$ is a Young diagram $\kappa$ with its boxes filled with $\lambda_1$ $1$'s, $\lambda_2$ $2$'s, etc. For a semi-standard $\kappa$-tableau of type $\lambda$, the labelling of the boxes is such that the rows are non-decreasing (from left to right) and the columns are strictly increasing (from top to bottom). E.g. |
<pre style="font-family: monospace;color:black"> | <pre style="font-family: monospace;color:black"> | ||
Line 30: | Line 40: | ||
is a semi-standard $(5,3,2)$-tableau of type $(4,2,2,2)$. The numbers $K(\kappa,\lambda)$ of semi-standard $\kappa$-tableaux of type $\lambda$ are called Kostka numbers. | is a semi-standard $(5,3,2)$-tableau of type $(4,2,2,2)$. The numbers $K(\kappa,\lambda)$ of semi-standard $\kappa$-tableaux of type $\lambda$ are called Kostka numbers. | ||
− | To each partition $\mu$ of $n$, there are associated two "natural" representations of $S_n$, the symmetric group on $n$ letters: the induced representation | + | To each partition $\mu$ of $n$, there are associated two "natural" representations of $S_n$, the symmetric group on $n$ letters: the induced representation $\rho(\mu)$ and the Specht module $[\mu]$. The representation $\rho(u)$ is: |
− | + | $$ | |
+ | \rho(\mu) = \operatorname{Ind}_{G_\mu}^{S_n} 1, | ||
+ | $$ | ||
− | where | + | where $1$ is the [[trivial representation]] of $G_\mu$ and $G_\mu$ is the Young subgroup of $S_n$ determined by $\mu$, $G_\mu = S_{\mu_1} \times \cdots \times S_{\mu_m}$, where $S_{\mu_i} = \{1\}$ if $\mu_i = 0$ and otherwise $S_{\mu_i}$ is the subgroup of permutations on the letters $\mu_1 + \dots + \mu_{i-1} + 1, \ldots, \mu_1 + \dots + \mu_i$. |
− | The group | + | The group $S_n$ acts on the set of all $\mu$-tableaux by permuting the labels. Two $\mu$-tableaux are equivalent if they differ by a permutation of their labels keeping the sets of indices in each row set-wise invariant. An equivalence class of $\mu$-tableaux is a $\mu$-tabloid. The action of $S_n$ on $\mu$-tableaux induces an action on $\mu$-tabloids, and extending this linearly over a base field $F$ gives a representation of $S_n$ which is evidently isomorphic to $\rho(\mu)$. The dimension of $\rho(\mu)$ is $(\square_\mu^n)$. Given a $\mu$-tableau $t$, let $\kappa_\tau$ be the following element of $F[S_n]$: |
− | + | $$ | |
+ | \kappa_t = \sum_{\pi \in C_t} \operatorname{sign}(\pi) \pi, | ||
+ | $$ | ||
− | where | + | where $C_t$ is the column-stabilizer of $t$, i.e. the subgroup of $S_n$ of all permutations that leave the labels of the columns of $t$ set-wise invariant. |
− | The Specht module, | + | The Specht module, $[\mu]$, of $\mu$ is the submodule of $\rho(\mu)$ spanned by all the elements $\kappa_t\{t\}$, where $\{t\}$ is the tabloid of $t$ and $t$ is a $\mu$-tableau. Over a field of characteristic zero the Specht modules give precisely all the different absolutely-irreducible representations of $S_n$. By Young's rule, the number of times that the Specht module $[\lambda]$ over $\mathbf{Q}$ occurs (as a composition factor) in $\rho(\kappa)$ is equal to the Kostka number $K(\kappa,\lambda)$. If $e_t$ is the [[Young symmetrizer]] of a $\mu$-tableau $t$, then the Specht module defined by the underlying diagram is isomorphic to the ideal $F[S_n]e_t$ of $F[S_n]$. This is also (up to isomorphism) the representation denoted by $T_{\mu}$ in [[Representation of the symmetric groups]]. Cf. [[Majorization ordering]] for a number of other results involving partitions, Young diagrams and tableaux, and representations of the symmetric groups. |
====References==== | ====References==== | ||
− | <table><TR><TD valign="top">[a1]</TD> <TD valign="top"> D. Knuth, "The art of computer programming" , '''3''' , Addison-Wesley (1973)</TD></TR></table> | + | <table> |
+ | <TR><TD valign="top">[a1]</TD> <TD valign="top"> D. Knuth, "The art of computer programming" , '''3''' , Addison-Wesley (1973)</TD></TR> | ||
+ | </table> | ||
− | {{TEX| | + | {{TEX|done}} |
Latest revision as of 03:36, 15 February 2024
of order $m$
A Young diagram of order $m$ in whose cells the different numbers $1,\ldots,m$ have been inserted in some order, e.g.
┌───┬───┬───┬───┐ │ 5 │ 7 │ 9 │ 4 │ ├───┼───┼───┼───┘ │ 8 │ 2 │ 1 │ ├───┼───┴───┘ │ 3 │ ├───┤ │ 6 │ └───┘
A Young tableau is called standard if in each row and in each column the numbers occur in increasing order. The number of all Young tableaux for a given Young diagram $t$ of order $m$ is equal to $m!$ and the number of standard Young tableaux is $$ \frac{m!}{\prod\lambda_{ij}} $$
where the product extends over all the cells $c_{ij}$ of $t$ and $\lambda_{ij}$ denotes the length of the corresponding hook.
Comments
In Western literature the phrase Ferrers diagram is also used for a Young diagram. In the Russian literature the phrase "Young tableau" ( "Yunga tablitsa" ) and "Young diagram" ( "Yunga diagramma" ) are used precisely in the opposite way, with "tablitsa" referring to the pictorial representation of a partition and "diagramma" being a filled-in "tablitsa" .
Let $\kappa$ denote a partition of $m$ ($\kappa=(\kappa_1,\ldots,\kappa_t)$, $\kappa_i \in \{0,1,\ldots\}$, $\kappa_1+\cdots+\kappa_t=m$) as well as its corresponding Young diagram, its pictorial representation. Let $\lambda$ be a second partition of $m$. A $\kappa$-tableau of type $\lambda$ is a Young diagram $\kappa$ with its boxes filled with $\lambda_1$ $1$'s, $\lambda_2$ $2$'s, etc. For a semi-standard $\kappa$-tableau of type $\lambda$, the labelling of the boxes is such that the rows are non-decreasing (from left to right) and the columns are strictly increasing (from top to bottom). E.g.
┌───┬───┬───┬───┬───┐ │ 1 │ 1 │ 1 │ 1 │ 4 │ ├───┼───┼───┼───┴───┘ │ 2 │ 2 │ 3 │ ├───┼───┼───┘ │ 3 │ 4 │ └───┴───┘
is a semi-standard $(5,3,2)$-tableau of type $(4,2,2,2)$. The numbers $K(\kappa,\lambda)$ of semi-standard $\kappa$-tableaux of type $\lambda$ are called Kostka numbers.
To each partition $\mu$ of $n$, there are associated two "natural" representations of $S_n$, the symmetric group on $n$ letters: the induced representation $\rho(\mu)$ and the Specht module $[\mu]$. The representation $\rho(u)$ is:
$$ \rho(\mu) = \operatorname{Ind}_{G_\mu}^{S_n} 1, $$
where $1$ is the trivial representation of $G_\mu$ and $G_\mu$ is the Young subgroup of $S_n$ determined by $\mu$, $G_\mu = S_{\mu_1} \times \cdots \times S_{\mu_m}$, where $S_{\mu_i} = \{1\}$ if $\mu_i = 0$ and otherwise $S_{\mu_i}$ is the subgroup of permutations on the letters $\mu_1 + \dots + \mu_{i-1} + 1, \ldots, \mu_1 + \dots + \mu_i$.
The group $S_n$ acts on the set of all $\mu$-tableaux by permuting the labels. Two $\mu$-tableaux are equivalent if they differ by a permutation of their labels keeping the sets of indices in each row set-wise invariant. An equivalence class of $\mu$-tableaux is a $\mu$-tabloid. The action of $S_n$ on $\mu$-tableaux induces an action on $\mu$-tabloids, and extending this linearly over a base field $F$ gives a representation of $S_n$ which is evidently isomorphic to $\rho(\mu)$. The dimension of $\rho(\mu)$ is $(\square_\mu^n)$. Given a $\mu$-tableau $t$, let $\kappa_\tau$ be the following element of $F[S_n]$:
$$ \kappa_t = \sum_{\pi \in C_t} \operatorname{sign}(\pi) \pi, $$
where $C_t$ is the column-stabilizer of $t$, i.e. the subgroup of $S_n$ of all permutations that leave the labels of the columns of $t$ set-wise invariant.
The Specht module, $[\mu]$, of $\mu$ is the submodule of $\rho(\mu)$ spanned by all the elements $\kappa_t\{t\}$, where $\{t\}$ is the tabloid of $t$ and $t$ is a $\mu$-tableau. Over a field of characteristic zero the Specht modules give precisely all the different absolutely-irreducible representations of $S_n$. By Young's rule, the number of times that the Specht module $[\lambda]$ over $\mathbf{Q}$ occurs (as a composition factor) in $\rho(\kappa)$ is equal to the Kostka number $K(\kappa,\lambda)$. If $e_t$ is the Young symmetrizer of a $\mu$-tableau $t$, then the Specht module defined by the underlying diagram is isomorphic to the ideal $F[S_n]e_t$ of $F[S_n]$. This is also (up to isomorphism) the representation denoted by $T_{\mu}$ in Representation of the symmetric groups. Cf. Majorization ordering for a number of other results involving partitions, Young diagrams and tableaux, and representations of the symmetric groups.
References
[a1] | D. Knuth, "The art of computer programming" , 3 , Addison-Wesley (1973) |
Young tableau. Encyclopedia of Mathematics. URL: http://encyclopediaofmath.org/index.php?title=Young_tableau&oldid=54276