Ecuaciones Diferenciales I: Teorema de existencia y unicidad para sistemas lineales

Por Omar González Franco

El estudio profundo de la naturaleza es la fuente
más fértil de descubrimientos matemáticos.
– Fourier

Introducción

Nos estamos acercando al final de la unidad 3 del curso.

En esta unidad estudiamos los sistemas lineales de primer orden con coeficientes constantes tanto homogéneos como no homogéneos, es por ello que en esta entrada demostraremos los teoremas de existencia y unicidad para estos sistemas lineales en particular.

Es interesante la enorme utilidad que tiene definir la exponencial de una matriz, este concepto nos ha permitido desarrollar distintos métodos de resolución y no solo ello, ahora nos permitirá demostrar estos teoremas de existencia y unicidad.

Comencemos con el caso homogéneo.

Teorema de existencia y unicidad para sistemas lineales homogéneos

Recordemos que estamos estudiando sistemas lineales homogéneo con coeficientes constantes, es decir, la matriz $\mathbf{A}$ es de la forma

$$\mathbf{A} = \begin{pmatrix}
a_{11} & a_{12} & \cdots & a_{1n} \\
a_{21} & a_{22} & \cdots & a_{2n} \\
\vdots & & & \vdots \\
a_{n1} & a_{n2} & \cdots & a_{nn}
\end{pmatrix} \label{1} \tag{1}$$

con $a_{ij}$, $i, j \in \{ 1, 2, 3, \cdots, n \}$ constantes.

Demostración: Recordemos que ya demostramos que la función (\ref{3}) es solución del sistema homogéneo para el caso $t_{0} = 0$, lo mismo podemos hacer para cualquier $t_{0}$, con ello estaremos mostrando la existencia. Consideremos la función (\ref{3}).

$$\mathbf{Y}(t) = e^{(t -t_{0}) \mathbf{A}} \mathbf{Y}_{0}$$

Con $t_{0} \in \mathbb{R}$. La derivada de esta función vectorial es

$$\mathbf{Y}^{\prime} = \mathbf{A} e^{(t -t_{0}) \mathbf{A}} \mathbf{Y}_{0} = \mathbf{A} \left( e^{(t -t_{0}) \mathbf{A}} \mathbf{Y}_{0} \right) = \mathbf{AY}$$

Esto muestra que es solución del sistema, veamos que efectivamente satisface los valores iniciales evaluando en $t_{0}$.

$$\mathbf{Y}(t_{0}) = e^{(t_{0} -t_{0}) \mathbf{A}} \mathbf{Y}_{0} = e^{\mathbf{0}} \mathbf{Y}_{0} = \mathbf{I} \cdot \mathbf{Y}_{0} = \mathbf{Y}_{0}$$

Con esto queda demostrado que la función $\mathbf{Y}(t) = e^{(t -t_{0})\mathbf{A}} \mathbf{Y}_{0}$ es solución del problema con valores iniciales, es decir, existe.

Para demostrar la unicidad supongamos que $\mathbf{X}(t)$ es otra solución del problema de valores iniciales. Consideremos una función derivable dada de la siguiente forma.

$$\mathbf{Z}(t) = e^{-(t -t_{0}) \mathbf{A}} \mathbf{X}(t) \label{4} \tag{4}$$

Si derivamos esta función obtenemos lo siguiente.

$$\mathbf{Z}^{\prime} = -\mathbf{A} e^{-(t -t_{0}) \mathbf{A}} \mathbf{X} + e^{-(t -t_{0}) \mathbf{A}} \mathbf{X}^{\prime} \label{5} \tag{5}$$

Como $\mathbf{X}$ es solución del sistema, entonces se satisface que $\mathbf{X}^{\prime} = \mathbf{AX}$. Sustituimos en la ecuación anterior.

\begin{align*}
\mathbf{Z}^{\prime} &= -\mathbf{A} e^{-(t -t_{0}) \mathbf{A}} \mathbf{X} + e^{-(t -t_{0}) \mathbf{A}} \left( \mathbf{AX} \right) \\
&= -\mathbf{A} e^{-(t -t_{0}) \mathbf{A}} \mathbf{X} + \left( e^{-(t -t_{0}) \mathbf{A}} \mathbf{A} \right) \mathbf{X} \\
&= -\mathbf{A} e^{-(t -t_{0}) \mathbf{A}} \mathbf{X} + \mathbf{A} e^{-(t -t_{0}) \mathbf{A}} \mathbf{X} \\
&= \mathbf{0}
\end{align*}

En donde hemos hecho uso de la propiedad

$$e^{-(t -t_{0}) \mathbf{A}} \mathbf{A} = \mathbf{A} e^{-(t -t_{0}) \mathbf{A}} \label{6} \tag{6}$$

Justifica este hecho.

El calculo anterior muestra que $\mathbf{Z}^{\prime}(t) = \mathbf{0}$, esto implica que $\mathbf{Z}(t)$ es una matriz constante. Evaluemos $t_{0}$ en (\ref{4}).

$$\mathbf{Z}(t_{0}) = e^{-(t_{0} -t_{0}) \mathbf{A}} \mathbf{X}(t_{0})$$

Como $\mathbf{X}$ satisface el problema de valores iniciales, entonces $\mathbf{X}(t_{0}) = \mathbf{Y}_{0}$. Sustituimos en el resultado anterior.

$$\mathbf{Z}(t_{0}) = e^{\mathbf{0}} \mathbf{Y}_{0} = \mathbf{I} \cdot \mathbf{Y}_{0} = \mathbf{Y}_{0} \label{7} \tag{7}$$

Hemos probado que $\mathbf{Z}$ es una matriz constante y que $\mathbf{Z}(t_{0}) = \mathbf{Y}_{0}$, entonces

$$\mathbf{Z}(t) = \mathbf{Y}_{0} \label{8} \tag{8}$$

para todo $t \in \mathbb{R}$. Sustituyendo (\ref{4}) en (\ref{8}), se tiene

$$e^{-(t -t_{0}) \mathbf{A}} \mathbf{X} = \mathbf{Y}_{0} \label{9} \tag{9}$$

Multipliquemos $e^{(t -t_{0}) \mathbf{A}}$ en ambos lados de esta ecuación por el lado izquierdo.

$$e^{(t -t_{0}) \mathbf{A}} e^{-(t -t_{0}) \mathbf{A}} \mathbf{X} = e^{(t -t_{0}) \mathbf{A}} \mathbf{Y}_{0} \label{10} \tag{10}$$

Por un lado,

$$e^{(t -t_{0}) \mathbf{A}} e^{-(t -t_{0}) \mathbf{A}} \mathbf{X} = \mathbf{I} \cdot \mathbf{X} = \mathbf{X}(t) \label{11} \tag{11}$$

Por otro lado, de (\ref{3}) se tiene que

$$e^{(t -t_{0}) \mathbf{A}} \mathbf{Y}_{0} = \mathbf{Y}(t) \label{12} \tag{12}$$

Entonces (\ref{10}) queda como

$$\mathbf{X}(t) = \mathbf{Y}(t) \label{13} \tag{13}$$

Es decir, la solución es única.

$\square$

Con este teorema queda justificado el método de valores y vectores propios que presentamos en esta unidad para resolver sistemas lineales homogéneos con coeficientes constantes.

Ahora demostremos el teorema de existencia y unicidad para el caso no homogéneo.

Teorema de existencia y unicidad para sistemas lineales no homogéneos

Demostración: Sea $\mathbf{Y}(t)$ una función arbitraría definida en un intervalo común $\delta$ que es solución del problema de valores iniciales (\ref{14}), entonces

$$\mathbf{Y}^{\prime} = \mathbf{AY} + \mathbf{G} \label{16} \tag{16}$$

y

$$\mathbf{Y}(t_{0}) = \mathbf{Y}_{0} \label{17} \tag{17}$$

Multipliquemos ambos lados de la ecuación (\ref{16}) por $e^{-t \mathbf{A}}$ por el lado izquierdo.

\begin{align*}
e^{-t \mathbf{A}} \mathbf{Y}^{\prime} &= e^{-t \mathbf{A}} \mathbf{AY} + e^{-t \mathbf{A}} \mathbf{G} \\
e^{-t \mathbf{A}} \mathbf{Y}^{\prime} -e^{-t \mathbf{A}} \mathbf{AY} &= e^{-t \mathbf{A}} \mathbf{G} \\
e^{-t \mathbf{A}} \mathbf{Y}^{\prime} -\mathbf{A} e^{-t \mathbf{A}} \mathbf{Y} &= e^{-t \mathbf{A}} \mathbf{G} \label{18} \tag{18}
\end{align*}

Notemos que

$$\left( e^{-t \mathbf{A}} \mathbf{Y} \right)^{\prime} = e^{-t \mathbf{A}} \mathbf{Y}^{\prime} -\mathbf{A} e^{-t \mathbf{A}} \mathbf{Y} \label{19} \tag{19}$$

Entonces el resultado (\ref{18}) se puede escribir como

$$\left( e^{-t \mathbf{A}} \mathbf{Y} \right)^{\prime} = e^{-t \mathbf{A}} \mathbf{G} \label{20} \tag{20}$$

Integremos de $t_{0}$ a $t$ y usemos el teorema fundamental del cálculo.

$$e^{-t \mathbf{A}} \mathbf{Y}(t) -e^{-t_{0} \mathbf{A}} \mathbf{Y}(t_{0}) = \int_{t_{0}}^{t}e^{-s \mathbf{A}} \mathbf{G}(s) ds \label{21} \tag{21}$$

Multipliquemos ambos lados de esta ecuación por $e^{t \mathbf{A}}$ por el lado izquierdo y sustituyamos los valores iniciales (\ref{17}).

\begin{align*}
e^{t \mathbf{A}} e^{-t \mathbf{A}} \mathbf{Y}(t) -e^{t \mathbf{A}} e^{-t_{0} \mathbf{A}} \mathbf{Y}_{0} &= e^{t \mathbf{A}} \int_{t_{0}}^{t} e^{-s \mathbf{A}} \mathbf{G}(s) ds \\
\mathbf{Y}(t) -e^{(t -t_{0}) \mathbf{A}} \mathbf{Y}_{0} &= e^{t \mathbf{A}} \int_{t_{0}}^{t} e^{-s \mathbf{A}} \mathbf{G}(s) ds
\end{align*}

Si despejamos a la función $\mathbf{Y}(t)$ obtenemos finalmente que

$$\mathbf{Y}(t) = e^{(t -t_{0}) \mathbf{A}} \mathbf{Y}_{0} + e^{t \mathbf{A}} \int_{t_{0}}^{t} e^{-s \mathbf{A}} \mathbf{G}(s) ds \label{22} \tag{22}$$

Como la función $\mathbf{Y}(t)$ fue arbitraría, con el proceso realizado estamos mostrando tanto la existencia como la unicidad de esta solución.

$\square$

Con este teorema queda justificado el método de variación de parámetros visto en la entada anterior.

Hemos concluido con esta entrada.

Tarea moral

Los siguientes ejercicios no forman parte de la evaluación del curso, pero servirán para entender mucho mejor los conceptos vistos en esta entrada, así como temas posteriores.

  1. Resolver los siguientes problemas con valores iniciales.
  • $\mathbf{Y}^{\prime}(t) = \begin{pmatrix}
    1 & 10 \\ 6 & -3
    \end{pmatrix} \mathbf{Y}(t), \hspace{1cm} \mathbf{Y}(0) = \begin{pmatrix}
    2 \\ 0
    \end{pmatrix}$
  • $\mathbf{Y}^{\prime}(t) = \begin{pmatrix}
    1 & 1 & 1 \\ 1 & 1 & 1 \\ -2 & -2 & -2
    \end{pmatrix} \mathbf{Y}(t), \hspace{1cm} \mathbf{Y}(0) =\begin{pmatrix}
    1 \\ -4 \\ 6
    \end{pmatrix}$
  1. Resolver los siguientes problemas con valores iniciales.
  • $\mathbf{Y}^{\prime}(t) = \begin{pmatrix}
    3 & -1 \\ -1 & 3
    \end{pmatrix} \mathbf{Y}(t) + \begin{pmatrix}
    4 e^{2t} \\ 4 e^{4t}
    \end{pmatrix}, \hspace{1cm} \mathbf{Y}(0) = \begin{pmatrix}
    1 \\ 1
    \end{pmatrix}$
  • $\mathbf{Y}^{\prime}(t) = \begin{pmatrix}
    1 & -1 \\ 1 & -1
    \end{pmatrix} \mathbf{Y}(t) + \begin{pmatrix}
    1/t \\ 1/t
    \end{pmatrix}, \hspace{1cm} \mathbf{Y}(1) = \begin{pmatrix}
    2 \\ -1
    \end{pmatrix}$

Más adelante…

En la siguiente y última entrada de esta unidad justificaremos el teorema de existencia y unicidad de los sistemas de ecuaciones diferenciales de primer orden lineales y no lineales. Es decir, el resultado general de los sistemas de ecuaciones diferenciales.

Entradas relacionadas

Agradecimientos

Trabajo realizado con el apoyo del Programa UNAM-DGAPA-PAPIME PE104522 «Hacia una modalidad a distancia de la Licenciatura en Matemáticas de la FC-UNAM – Etapa 2»

Deja una respuesta

Tu dirección de correo electrónico no será publicada. Los campos obligatorios están marcados con *

Este sitio usa Akismet para reducir el spam. Aprende cómo se procesan los datos de tus comentarios.