aboutsummaryrefslogtreecommitdiffstats
path: root/buch/chapters/40-eigenwerte/uebungsaufgaben/4003.tex
diff options
context:
space:
mode:
authorAndreas Müller <andreas.mueller@othello.ch>2021-01-18 21:01:26 +0100
committerAndreas Müller <andreas.mueller@othello.ch>2021-01-18 21:01:26 +0100
commit812a80acf52275699afb285db46aa76be03006c2 (patch)
treed603bf9c54cd29909551d1cb7f5fd7b66c64ca4e /buch/chapters/40-eigenwerte/uebungsaufgaben/4003.tex
parentneue Sachen zur linearen Algebra (diff)
downloadSeminarMatrizen-812a80acf52275699afb285db46aa76be03006c2.tar.gz
SeminarMatrizen-812a80acf52275699afb285db46aa76be03006c2.zip
add more stuff about spectral radius
Diffstat (limited to '')
-rw-r--r--buch/chapters/40-eigenwerte/uebungsaufgaben/4003.tex241
1 files changed, 241 insertions, 0 deletions
diff --git a/buch/chapters/40-eigenwerte/uebungsaufgaben/4003.tex b/buch/chapters/40-eigenwerte/uebungsaufgaben/4003.tex
new file mode 100644
index 0000000..3cd9959
--- /dev/null
+++ b/buch/chapters/40-eigenwerte/uebungsaufgaben/4003.tex
@@ -0,0 +1,241 @@
+Finden Sie eine Basis von $\mathbb{Q}^4$ derart, dass die Matrix $A$
+\[
+A
+=
+\begin{pmatrix}
+-13& 5& -29& 29\\
+-27& 11& -51& 51\\
+ -3& 1& -2& 5\\
+ -6& 2& -10& 13
+\end{pmatrix}
+\]
+Jordansche Normalform hat.
+
+\begin{loesung}
+Zunächst muss man die Eigenwerte finden.
+Dazu kann man das charakteristische Polynom berechnen, man findet nach
+einiger Rechnung oder mit Hilfe einer Software für symbolische Rechnung:
+\[
+\chi_A(\lambda)
+=
+x^4-9x^3+30x^2-44x+24
+=
+(x-3)^3(x-2),
+\]
+Eigenwerte sind also $\lambda=3$ und $\lambda=2$.
+
+Der Eigenwert $\lambda=2$ ist ein einfacher Eigenwert, der zugehörige
+Eigenraum ist daher eindimensional.
+Ein Eigenvektor kann mit Hilfe des linearen Gleichungssystems
+\begin{align*}
+\begin{tabular}{|>{$}c<{$}>{$}c<{$}>{$}c<{$}>{$}c<{$}|}
+\hline
+-13-\lambda& 5 &-29 &29 \\
+-27 &11-\lambda&-51 &51 \\
+ -3 & 1 & -2-\lambda& 5 \\
+ -6 & 2 &-10 &13-\lambda\\
+\hline
+\end{tabular}
+&\rightarrow
+\begin{tabular}{|>{$}c<{$}>{$}c<{$}>{$}c<{$}>{$}c<{$}|}
+\hline
+ -16& 5& -29& 29\\
+ -27& 8& -51& 51\\
+ -3& 1& -5& 5\\
+ -6& 2& -10& 10\\
+\hline
+\end{tabular}
+\to
+\begin{tabular}{|>{$}c<{$}>{$}c<{$}>{$}c<{$}>{$}c<{$}|}
+\hline
+1&0&0& 0\\
+0&1&0& 0\\
+0&0&1&-1\\
+0&0&0& 0\\
+\hline
+\end{tabular}
+\end{align*}
+gefunden werden.
+Daraus liest man den Eigenvektor
+\[
+b_1
+=
+\begin{pmatrix} 0\\0\\1\\1\end{pmatrix},
+\qquad
+Ab_1 =
+\begin{pmatrix}
+-13& 5& -29& 29\\
+-27& 11& -51& 51\\
+ -3& 1& -2& 5\\
+ -6& 2& -10& 13
+\end{pmatrix}
+\begin{pmatrix} 0\\0\\1\\1\end{pmatrix}
+=
+\begin{pmatrix}
+0\\0\\3\\3
+\end{pmatrix}
+=
+3b_1
+\]
+ab.
+Diesen Vektor können wir auch finden, indem wir $\mathcal{J}(A-eE)$
+bestimmen.
+Die vierte Potenz von $A-2E$ ist
+\begin{equation}
+(A-2E)^4
+=
+\begin{pmatrix}
+ 0& 0& 0& 0\\
+ 0& 0& 0& 0\\
+ 0& 0& 2& -1\\
+ 0& 0& 2& -1
+\end{pmatrix},
+\label{4003:potenz}
+\end{equation}
+der zugehörige Bildraum ist wieder aufgespannt von $b_1$.
+
+Aus \eqref{4003:potenz} kann man aber auch eine Basis
+\[
+b_2
+=
+\begin{pmatrix}1\\0\\0\\0\end{pmatrix}
+,\qquad
+b_3
+=
+\begin{pmatrix}0\\1\\0\\0\end{pmatrix}
+,\qquad
+b_4
+=
+\begin{pmatrix}0\\0\\1\\2\end{pmatrix}
+\]
+für den Kern $\mathcal{K}(A-2E)$ ablesen.
+Da $\lambda=2$ der einzige andere Eigenwert ist, muss $\mathcal{K}(A-2E)
+= \mathcal{J}(A-3E)$ sein.
+Dies lässt sich überprüfen, indem wir die vierte Potenz von $A-2E$
+berechnen, sie ist
+\[
+(A-2E)^4
+=
+\begin{pmatrix}
+ 79& -26& 152& -152\\
+ 162& -53& 312& -312\\
+ 12& -4& 23& -23\\
+ 24& -8& 46& -46\\
+\end{pmatrix}.
+\]
+Die Spaltenvektoren lassen sich alle durch die Vektoren $b_2$, $b_3$
+und $b_4$ ausdrücken, also ist $\mathcal{J}(A-2E)=\langle b_2,b_3,b_4\rangle$.
+
+Indem die Vektoren $b_i$ als Spalten in eine Matrix $T$ schreibt, kann man
+jetzt berechnen, wie die Matrix der linearen Abbildung in dieser neuen
+Basis aussieht, es ist
+\[
+A'=T^{-1}AT
+\left(
+\begin{array}{r|rrr}
+ 3& 0& 0& 0\\
+\hline
+ 0& -13& 5& 29\\
+ 0& -27& 11& 51\\
+ 0& -3& 1& 8
+\end{array}
+\right),
+\]
+wir haben also tatsächlich die versprochene Blockstruktur.
+
+Der $3\times 3$-Block
+\[
+A_1
+=
+\begin{pmatrix}
+ -13& 5& 29\\
+ -27& 11& 51\\
+ -3& 1& 8
+\end{pmatrix}
+\]
+in der rechten unteren Ecke hat den dreifachen Eigenwert $2$,
+und die Potenzen von $A_1-2E$ sind
+\[
+A_1-2E
+\begin{pmatrix}
+ -15 & 5& 29\\
+ -27 & 9& 51\\
+ -3 & 1& 6
+\end{pmatrix}
+,\qquad
+(A_1-2E)^2
+=
+\begin{pmatrix}
+ 3 & -1 & -6\\
+ 9 & -3 &-18\\
+ 0 & 0 & 0\\
+\end{pmatrix}
+,\qquad
+(A_1-2E)^3=0.
+\]
+Für die Jordan-Normalform brauchen wir einen von $0$ verschiedenen
+Vektor im Kern von $(A_1-2E)^2$, zum Beispiel den Vektor mit den
+Komponenten $1,3,1$.
+Man beachte aber, dass diese Komponenten jetzt in der neuen Basis
+$b_2,\dots,b_4$ zu verstehen sind, d.~h.~der Vektor, den wir suchen, ist
+\[
+c_3
+=
+b_1+ 3b_2+b_3
+=
+\begin{pmatrix}1\\3\\1\\2\end{pmatrix}.
+\]
+Jetzt berechnen wir die Bilder von $c_3$ unter $A-2E$:
+\[
+c_2
+=
+\begin{pmatrix}
+29\\51Ò\\6\\12
+\end{pmatrix}
+,\qquad
+c_1
+=
+\begin{pmatrix}
+-6\\-18\\0\\0
+\end{pmatrix}.
+\]
+Die Basis $b_1,c_1,c_2,c_3$ ist also eine Basis, in der die Matrix $A$
+Jordansche Normalform annimmt.
+
+Die Umrechnung der Matrix $A$ in die Basis $\{b_1,c_1,c_2,c_3\}$ kann
+mit der Matrix
+\[
+T_1
+=
+\begin{pmatrix}
+ 0& -6& 29& 1\\
+ 0& -18& 51& 3\\
+ 1& 0& 6& 1\\
+ 1& 0& 12& 2\\
+\end{pmatrix},
+\qquad
+T_1^{-1}
+=
+\frac{1}{216}
+\begin{pmatrix}
+ 0& 0& 432& -216\\
+ 33& -23& -36& 36\\
+ 18& -6& 0& 0\\
+ -108& 36& -216& 216
+\end{pmatrix}
+\]
+erfolgen und ergibt die Jordansche Normalform
+\[
+A'
+=
+\begin{pmatrix}
+3&0&0&0\\
+0&2&1&0\\
+0&0&2&1\\
+0&0&0&2
+\end{pmatrix}
+\]
+wie erwartet.
+\end{loesung}
+
+