[skip-to-content]
\(\newcommand{\identity}{\mathrm{id}} \newcommand{\notdivide}{{\not{\mid}}} \newcommand{\notsubset}{\not\subset} \newcommand{\lcm}{\operatorname{lcm}} \newcommand{\gf}{\operatorname{GF}} \newcommand{\inn}{\operatorname{Inn}} \newcommand{\aut}{\operatorname{Aut}} \newcommand{\Hom}{\operatorname{Hom}} \newcommand{\cis}{\operatorname{cis}} \newcommand{\chr}{\operatorname{char}} \newcommand{\Null}{\operatorname{Null}} \renewcommand{\gcd}{\operatorname{mcd}} \renewcommand{\lcm}{\operatorname{mcm}} \renewcommand{\deg}{\operatorname{gr}} \newcommand{\lt}{<} \newcommand{\gt}{>} \newcommand{\amp}{&} \)

Sección12.1Grupos de Matrices

SubsecciónAlgunos Resultados de Álgebra Lineal

Antes de estudiar grupos de matrices, debemos ercordar algunos resultados básicos de álgebra lineal. Una de las ideas fundamentales de álgebra lineal es la de una transformación lineal. Una transformación lineal o función lineal \(T : {\mathbb R}^n \rightarrow {\mathbb R}^m\) es una función que respeta (o preserva) la suma de vectores y la multiplicación por escalares; es decir, para vectores \({\mathbf x}\) e \({\mathbf y}\) en \({\mathbb R}^n\) y un escalar \(\alpha \in {\mathbb R}\text{,}\)

\begin{align*} T({\mathbf x}+{\mathbf y}) & = T({\mathbf x}) + T({\mathbf y})\\ T(\alpha {\mathbf y}) & = \alpha T({\mathbf y}). \end{align*}

Una matriz de \(m \times n\) con coeficientes en \({\mathbb R}\) representa una transformación lineal de \({\mathbb R}^n\) a \({\mathbb R}^m\text{.}\) Si escribimos vectores \({\mathbf x} = (x_1, \ldots, x_n)^{\rm t}\) e \({\mathbf y} = (y_1, \ldots, y_n)^{\rm t}\) en \({\mathbb R}^n\) como matrices de una columna, entonces una matriz de \(m \times n\)

\begin{equation*} A = \begin{pmatrix} a_{11} & a_{12} & \cdots & a_{1n} \\ a_{21} & a_{22} & \cdots & a_{2n} \\ \vdots & \vdots & \ddots & \vdots \\ a_{m1} & a_{m2} & \cdots & a_{mn} \end{pmatrix} \end{equation*}

envía a los vectores en \({\mathbb R}^m\) linealmente por multiplicación matricial. Observe que si \(\alpha\) es un número real,

\begin{equation*} A({\mathbf x} + {\mathbf y} ) = A {\mathbf x }+ A {\mathbf y} \qquad \text{and} \qquad \alpha A {\mathbf x} = A ( \alpha {\mathbf x}), \end{equation*}

donde

\begin{equation*} {\mathbf x} = \begin{pmatrix} x_1 \\ x_2 \\ \vdots \\ x_n \end{pmatrix}. \end{equation*}

Con frecuencia abreviaremos la matriz \(A\) escribiendo \((a_{ij})\text{.}\)

Recíprocamente, si \(T : {\mathbb R}^n \rightarrow {\mathbb R}^m\) es una función lineal, podemos asociar una matriz \(A\) con \(T\) considerando lo que \(T\) le hace a los vectores

\begin{align*} {\mathbf e}_1 & = (1, 0, \ldots, 0)^{\rm t}\\ {\mathbf e}_2 & = (0, 1, \ldots, 0)^{\rm t}\\ & \vdots & \\ {\mathbf e}_n & = (0, 0, \ldots, 1)^{\rm t}. \end{align*}

Podemos escribir cualquier vector \({\mathbf x} = (x_1, \ldots, x_n)^{\rm t}\) como

\begin{equation*} x_1 {\mathbf e}_1 + x_2 {\mathbf e}_2 + \cdots + x_n {\mathbf e}_n. \end{equation*}

Así, si

\begin{align*} T({\mathbf e}_1) & = (a_{11}, a_{21}, \ldots, a_{m1})^{\rm t},\\ T({\mathbf e}_2) & = (a_{12}, a_{22}, \ldots, a_{m2})^{\rm t},\\ & \vdots & \\ T({\mathbf e}_n) & = (a_{1n}, a_{2n}, \ldots, a_{mn})^{\rm t}, \end{align*}

entonces

\begin{align*} T({\mathbf x} ) & = T(x_1 {\mathbf e}_1 + x_2 {\mathbf e}_2 + \cdots + x_n {\mathbf e}_n)\\ & = x_1 T({\mathbf e}_1) + x_2 T({\mathbf e}_2) + \cdots + x_n T({\mathbf e}_n)\\ & = \left( \sum_{k=1}^{n} a_{1k} x_k, \ldots, \sum_{k=1}^{n} a_{mk} x_k \right)^{\rm t}\\ & = A {\mathbf x}. \end{align*}
Ejemplo12.1

Si \(T : {\mathbb R}^2 \rightarrow {\mathbb R}^2\) es la función dada por

\begin{equation*} T(x_1, x_2) = (2 x_1 + 5 x_2, - 4 x_1 + 3 x_2), \end{equation*}

los axiomas que \(T\) debe satisfacer para ser una transformación lineal se verifican fácilmente. Los vectores columna \(T {\mathbf e}_1 = (2, -4)^{\rm t}\) y \(T {\mathbf e}_2 = (5,3)^{\rm t}\) nos dicen que \(T\) está dada por la matriz

\begin{equation*} A = \begin{pmatrix} 2 & 5 \\ -4 & 3 \end{pmatrix}. \end{equation*}

Como estamos interesados en grupos de matrices, necesitamos saber qué matrices tienen inverso multiplicativo. Recuerde que una matriz \(A\) de \(n \times n\) es invertible si y solo si existe otra matriz \(A^{-1}\) tal que \(A A^{-1} = A^{-1} A = I\text{,}\) donde

\begin{equation*} I = \begin{pmatrix} 1 & 0 & \cdots & 0 \\ 0 & 1 & \cdots & 0 \\ \vdots & \vdots & \ddots & \vdots \\ 0 & 0 & \cdots & 1 \end{pmatrix} \end{equation*}

es la matriz identidad de \(n \times n\text{.}\) De álgebra lineal sabemos que \(A\) es invertible si y solo si el determinante de \(A\) es distinto de cero. También se dice que una matriz invertible es no singular.

Ejemplo12.2

Si \(A\) es la matriz

\begin{equation*} \begin{pmatrix} 2 & 1 \\ 5 & 3 \end{pmatrix}, \end{equation*}

entonces la inversa de \(A\) es

\begin{equation*} A^{-1} = \begin{pmatrix} 3 & -1 \\ -5 & 2 \end{pmatrix}. \end{equation*}

Sabemos que \(A^{-1}\) existe, pues \(\det(A) = 2 \cdot 3 - 5 \cdot 1 = 1\) no es cero.

Algunos otros hechos sobre determinantes resultarán útiles en el transcurso de este capítulo. Sean \(A\) y \(B\) matrices de \(n \times n\text{.}\) De álgebra lineal tenemos las siguientes propiedades de los determinantes.

  • El determinante es un homomorfismo al grupo multiplicativo de los números reales; es decir, \(\det( A B) = (\det A )(\det B)\text{.}\)

  • Si \(A\) es una matriz invertible, entonces \(\det(A^{-1}) = 1 / \det A\text{.}\)

  • Si definimos la transpuesta de una matriz \(A = (a_{ij})\) como \(A^{\rm t} = (a_{ji})\text{,}\) entonces \(\det(A^{\rm t}) = \det A\text{.}\)

  • Sea \(T\) la transformación lineal asociada con una matriz \(A\) de \(n \times n\text{.}\) Entonces \(T\) multiplica volúmenes por un facor de \(|\det A|\text{.}\) En el caso de \({\mathbb R}^2\text{,}\) esto quiere decir que \(T\) multiplica áreas por \(|\det A|\text{.}\)

Funciones lineales, matrices, y determinantes se pasan en un curso elemental de álgebra lineal; pero, si no ha tenido un curso así, es un proceso simple verificar estas propiedades directamente para matrices de \(2 \times 2\text{,}\) que es el caso que más nos interesará.

SubsecciónEl Grupo Lineal General y el Grupo Lineal Especial

El conjunto de todas las matrices invertibles de \(n \times n\) forma un grupo llamado grupo lineal general. Denotaremos este grupo por\(GL_n({\mathbb R})\text{.}\) El grupo lineal general tiene varios subgrupos importantes. La propiedad multiplicativa del determinante implica que el conjunto de las matrices cuyo determinante es uno es un subgrupo del grupo lineal general. Dicho de otra forma, supongamos que \(\det(A) =1\) y que \(\det(B) = 1\text{.}\) Entonces \(\det(AB) = \det(A) \det (B) = 1\) y \(\det(A^{-1}) = 1 / \det A = 1\text{.}\) Este subgrupo se llama grupo lineal especial y se denota por \(SL_n({\mathbb R})\text{.}\)

Ejemplo12.3

Dada una matriz de \(2 \times 2\)

\begin{equation*} A = \begin{pmatrix} a & b \\ c & d \end{pmatrix}, \end{equation*}

el determinante de \(A\) es \(ad-bc\text{.}\) El grupo \(GL_2({\mathbb R})\) consiste de aquellas matrices para las que \(ad-bc \neq 0\text{.}\) La inversa de \(A\) es

\begin{equation*} A^{-1} = \frac{1}{ad-bc} \begin{pmatrix} d & -b \\ -c & a \end{pmatrix}. \end{equation*}

Si \(A\) está en \(SL_2({\mathbb R})\text{,}\) entonces

\begin{equation*} A^{-1} = \begin{pmatrix} d & -b \\ -c & a \end{pmatrix}. \end{equation*}

Geométricamente, \(SL_2({\mathbb R})\) es el grupo que preserva las áreas de los paralelógramos. Sea

\begin{equation*} A = \begin{pmatrix} 1 & 1 \\ 0 & 1 \end{pmatrix} \end{equation*}

en \(SL_2({\mathbb R})\text{.}\) En la Figura 12.4, el cuadrado unitario correspondiente a los vectores \({\mathbf x} = (1,0)^{\rm t}\) y \({\mathbf y} = (0,1)^{\rm t}\) es enviado por \(A\) al paralelógramo con lados \((1,0)^{\rm t}\) y \((1, 1)^{\rm t}\text{;}\) es decir, \(A {\mathbf x} = (1,0)^{\rm t}\) y \(A {\mathbf y} = (1, 1)^{\rm t}\text{.}\) Note que estos dos paralelógramos tienen la misma área.

<<SVG image is unavailable, or your browser cannot render it>>

Figura12.4\(SL_2(\mathbb R)\) actuando en el cuadrado unitario

SubsecciónEl Grupo Ortogonal \(O(n)\)

Otro subgrupo de \(GL_n({\mathbb R})\) es el grupo ortogonal. Una matriz \(A\) es ortogonal si \(A^{-1} = A^{\rm t}\text{.}\) El grupo ortogonal consiste en el conjunto de todas las matrices ortogonales. Escribimos \(O(n)\) para el grupo ortogonal de \(n \times n\text{.}\) Dejamos como ejercicio demostrar que \(O(n)\) es un subgrupo de \(GL_n( {\mathbb R})\text{.}\)

Ejemplo12.5

Las siguiente matrices son ortogonales:

\begin{equation*} \begin{pmatrix} 3/5 & -4/5 \\ 4/5 & 3/5 \end{pmatrix}, \quad \begin{pmatrix} 1/2 & -\sqrt{3}/2 \\ \sqrt{3}/2 & 1/2 \end{pmatrix}, \quad \begin{pmatrix} -1/\sqrt{2} & 0 & 1/ \sqrt{2} \\ 1/\sqrt{6} & -2/\sqrt{6} & 1/\sqrt{6} \\ 1/ \sqrt{3} & 1/ \sqrt{3} & 1/ \sqrt{3} \end{pmatrix}. \end{equation*}

Hay una forma más geométrica de ver el grupo \(O(n)\text{.}\) Las matrices ortogonales son exactamente aquellas que preservan lad longitudes de los vectores. Podemos definir la longitud de un vector usando el producto interno Euclideano, o producto punto, de dos vectores. El producto interno Euclideano de dos vectores \({\mathbf x}=(x_1, \ldots, x_n)^{\rm t}\) y \({\mathbf y}=(y_1, \ldots, y_n)^{\rm t}\) es

\begin{equation*} \langle {\mathbf x}, {\mathbf y} \rangle = {\mathbf x}^{\rm t} {\mathbf y} = (x_1, x_2, \ldots, x_n) \begin{pmatrix} y_1 \\ y_2 \\ \vdots \\ y_n \end{pmatrix} = x_1 y_1 + \cdots + x_n y_n. \end{equation*}

Definimos la longitud de un vector \({\mathbf x}=(x_1, \ldots, x_n)^{\rm t}\) como

\begin{equation*} \| {\mathbf x} \| = \sqrt{\langle {\mathbf x}, {\mathbf x} \rangle} = \sqrt{x_1^2 + \cdots + x_n^2}. \end{equation*}

Asociada a la noción de longitud de un vector está la idea de distancia entre dos vectores. Definimos la distancia entre dos vectores \({\mathbf x}\) e \({\mathbf y}\) como \(\| {\mathbf x}-{\mathbf y} \|\text{.}\) Dejamos como ejercicio demostrar la siguiente proposición sobre las propiedades de los productos internos Euclideanos.

Ejemplo12.7

El vector \({\mathbf x} =(3,4)^{\rm t}\) tiene longitud \(\sqrt{3^2 + 4^2} = 5\text{.}\) Podemos también ver que la matriz ortogonal

\begin{equation*} A= \begin{pmatrix} 3/5 & -4/5 \\ 4/5 & 3/5 \end{pmatrix} \end{equation*}

preserva la longitud de este vector. El vector \(A{\mathbf x} = (-7/5,24/5)^{\rm t}\) también tiene longitud 5.

Como \(\det(A A^{\rm t}) = \det(I) = 1\) y \(\det(A) = \det( A^{\rm t} )\text{,}\) el determinante de cualquier matriz ortogonal es 1 o \(-1\text{.}\) Considere los vectores columna

\begin{equation*} {\mathbf a}_j = \begin{pmatrix} a_{1j} \\ a_{2j} \\ \vdots \\ a_{nj} \end{pmatrix} \end{equation*}

de la matriz ortogonal \(A= (a_{ij})\text{.}\) Since \(AA^{\rm t} = I\text{,}\) \(\langle {\mathbf a}_r, {\mathbf a}_s \rangle = \delta_{rs}\text{,}\) donde

\begin{equation*} \delta_{rs} = \left\{ \begin{array}{cc} 1 & r = s \\ 0 & r \neq s \end{array} \right. \end{equation*}

es la delta de Kronecker . Así, los vectores columna de una matriz ortogonal todos tienen longitud 1; y el producto interno Euclideano de vectores columna distintos es cero. Cualquier conjunto de vectores que satisface esta propiedad se llama conjunto ortonormal. Recíprocamente, dada una matriz \(A\) de \(n \times n\) cuyas columnas forman un conjunto ortonormal, se tiene que \(A^{-1} = A^{\rm t}\text{.}\)

Decimos que una matriz \(A\) preserva distancias, o preserva el producto interno cuando \(\| T{\mathbf x}- T{\mathbf y} \| =\| {\mathbf x}- {\mathbf y} \|\text{,}\) \(\| T{\mathbf x} \| =\| {\mathbf x} \|\text{,}\) o \(\langle T{\mathbf x}, T{\mathbf y} \rangle = \langle {\mathbf x},{\mathbf y} \rangle\text{,}\) respectivamente. El siguiente teorema, que caracteriza el grupo ortogonal, establece que estos conceptos son iguales.

Ya hemos mostrado la equivalencia de (1) y (2).

\((2) \Rightarrow (3)\text{.}\)

\begin{align*} \langle A{\mathbf x}, A{\mathbf y} \rangle & = (A {\mathbf x})^{\rm t} A {\mathbf y}\\ & = {\mathbf x}^{\rm t} A^{\rm t} A {\mathbf y}\\ & = {\mathbf x}^{\rm t} {\mathbf y}\\ & = \langle {\mathbf x}, {\mathbf y} \rangle. \end{align*}

\((3) \Rightarrow (2)\text{.}\) Como

\begin{align*} \langle {\mathbf x}, {\mathbf x} \rangle & = \langle A{\mathbf x}, A{\mathbf x} \rangle\\ & = {\mathbf x}^{\rm t} A^{\rm t} A {\mathbf x}\\ & = \langle {\mathbf x}, A^{\rm t} A{\mathbf x} \rangle, \end{align*}

sabemos que \(\langle {\mathbf x}, (A^{\rm t} A - I){\mathbf x} \rangle = 0\) para todo \({\mathbf x}\text{.}\) Por lo tanto, \(A^{\rm t} A -I = 0\) o \(A^{-1} = A^{\rm t}\text{.}\)

\((3) \Rightarrow (4)\text{.}\) Si \(A\) preserva el producto interno, entonces \(A\) preserva distancias, pues

\begin{align*} \| A{\mathbf x} - A{\mathbf y} \|^2 & = \| A({\mathbf x} - {\mathbf y}) \|^2\\ & = \langle A({\mathbf x} - {\mathbf y}), A({\mathbf x} - {\mathbf y}) \rangle\\ & = \langle {\mathbf x} - {\mathbf y}, {\mathbf x} - {\mathbf y} \rangle\\ & = \| {\mathbf x} - {\mathbf y} \|^2. \end{align*}

\((4) \Rightarrow (5)\text{.}\) Si \(A\) preserva distancias, entonces \(A\) preserva longitudes. Tomando \({\mathbf y} = 0\text{,}\) tenemos

\begin{equation*} \| A{\mathbf x}\| = \| A{\mathbf x}- A{\mathbf y} \| = \| {\mathbf x}- {\mathbf y} \| = \| {\mathbf x} \|. \end{equation*}

\((5) \Rightarrow (3)\text{.}\) Usamos la siguiente identidad para mostrar que la preservación de longitudes implica la preservación del producto interno:

\begin{equation*} \langle {\mathbf x}, {\mathbf y} \rangle = \frac{1}{2} \left[ \|{\mathbf x} +{\mathbf y}\|^2 - \|{\mathbf x}\|^2 - \|{\mathbf y}\|^2 \right]. \end{equation*}

Observe que

\begin{align*} \langle A {\mathbf x}, A {\mathbf y} \rangle & = \frac{1}{2} \left[ \|A {\mathbf x} + A {\mathbf y} \|^2 - \|A {\mathbf x} \|^2 - \|A {\mathbf y} \|^2 \right]\\ & = \frac{1}{2} \left[ \|A ( {\mathbf x} + {\mathbf y} ) \|^2 - \|A {\mathbf x} \|^2 - \|A {\mathbf y} \|^2 \right]\\ & = \frac{1}{2} \left[ \|{\mathbf x} + {\mathbf y}\|^2 - \|{\mathbf x}\|^2 - \|{\mathbf y}\|^2 \right]\\ & = \langle {\mathbf x}, {\mathbf y} \rangle. \end{align*}

<<SVG image is unavailable, or your browser cannot render it>>

Figura12.9\(O(2)\) actuando en \(\mathbb R^2\)
Ejemplo12.10

Examinemos el grupo ortogonal en\({\mathbb R}^2\) en mayor detalle. Un elemento \(T \in O(2)\) está determinado por su acción en \({\mathbf e}_1 = (1, 0)^{\rm t}\) y \({\mathbf e}_2 = (0, 1)^{\rm t}\text{.}\) Si \(T({\mathbf e}_1) = (a,b)^{\rm t}\text{,}\) entonces \(a^2 + b^2 = 1\) y \(T({\mathbf e}_2) = (-b, a)^{\rm t}\text{.}\) Luego, \(T\) puede ser representada por

\begin{equation*} A = \begin{pmatrix} a & -b \\ b & a \end{pmatrix} = \begin{pmatrix} \cos \theta & - \sin \theta \\ \sin \theta & \cos \theta \end{pmatrix}, \end{equation*}

donde \(0 \leq \theta \lt 2 \pi\text{.}\) Una matriz \(T\) en \(O(2)\) ya sea refleja o rota un vector en \({\mathbb R}^2\) (Figura 12.9). Una reflexión respecto al eje horizontal está dada por la matriz

\begin{equation*} \begin{pmatrix} 1 & 0 \\ 0 & -1 \end{pmatrix}, \end{equation*}

mientras una rotación en un ángulo \(\theta\) en sentido antihorario debe venir de una matriz de la forma

\begin{equation*} \begin{pmatrix} \cos \theta & \sin \theta \\ \sin \theta & -\cos \theta \end{pmatrix}. \end{equation*}

Una reflexión respecto a una recta \(\ell\) es simplemente una relfexión respecto al eje horizontal seguida de una rotación. Si \(\det A =-1\text{,}\) entonces \(A\) corresponde a una reflexión.

Dos de los otros grupos de matrices o relacionados a matrices que consideraremos son el grupo ortogonal especial y el grupo de movimientos Euclideanos. El grupo ortogonal especial, \(SO(n)\text{,}\) e simplemente la intersección de \(O(n)\) y \(SL_n({\mathbb R})\text{;}\) es decir, aquellos elementos en \(O(n)\) con determinante uno. El grupo Euclideano, \(E(n)\text{,}\) puede ser escrito como pares ordenados \((A, {\mathbf x})\text{,}\) donde \(A\) está en \(O(n)\) y \({\mathbf x}\) está en \({\mathbb R}^n\text{.}\) Definimos la multiplicación como

\begin{equation*} (A, {\mathbf x}) (B, {\mathbf y}) = (AB, A {\mathbf y} +{\mathbf x}). \end{equation*}

La identidad del grupo es \((I,{\mathbf 0})\text{;}\) el inverso de \((A, {\mathbf x})\) es \((A^{-1}, -A^{-1} {\mathbf x})\text{.}\) En el Ejercicio 12.3.6, debe verificar que \(E(n)\) es realmente un grupo con esta operación.

<<SVG image is unavailable, or your browser cannot render it>>

Figura12.11Traslaciones en \(\mathbb R^2\)