Aufgabe (Rechnen mit Polynomen, 5 Punkte)
Bearbeiten
Für
z
=
x
+
i
y
{\displaystyle z=x+iy}
ist
z
¯
=
x
−
i
y
{\displaystyle {\bar {z}}=x-iy}
. Lösen wir das Gleichungssystem
z
=
x
+
i
y
z
¯
=
x
−
i
y
{\displaystyle {\begin{array}{rl}z&=x+iy\\{\bar {z}}&=x-iy\end{array}}}
nach
x
,
y
{\displaystyle x,y}
auf, erhalten wir
x
=
1
2
(
z
+
z
¯
)
{\displaystyle x={\frac {1}{2}}(z+{\bar {z}})}
und
y
=
1
2
i
(
z
−
z
¯
)
{\displaystyle y={\frac {1}{2i}}(z-{\bar {z}})}
. Dies setzen wir in
p
{\displaystyle p}
ein, mit dem Binomialtheorem folgt
p
(
z
)
=
∑
κ
,
λ
a
κ
λ
x
κ
y
λ
=
∑
κ
,
λ
a
κ
λ
1
2
κ
(
z
+
z
¯
)
κ
1
(
2
i
)
λ
(
z
−
z
¯
)
λ
=
∑
κ
,
λ
a
κ
λ
2
κ
+
λ
i
λ
∑
α
=
0
κ
(
κ
α
)
z
α
z
¯
κ
−
α
⋅
∑
β
=
0
λ
(
−
1
)
λ
−
β
(
λ
β
)
z
β
z
¯
λ
−
β
=
∑
κ
,
λ
∑
α
=
0
κ
∑
β
=
0
λ
a
κ
λ
2
κ
+
λ
i
λ
(
κ
α
)
(
−
1
)
λ
−
β
(
λ
β
)
z
α
+
β
z
¯
κ
+
λ
−
(
α
+
β
)
{\displaystyle {\begin{array}{rl}p(z)&=\displaystyle \sum _{\kappa ,\lambda }a_{\kappa \lambda }x^{\kappa }y^{\lambda }\\&=\displaystyle \sum _{\kappa ,\lambda }a_{\kappa \lambda }{\frac {1}{2^{\kappa }}}(z+{\bar {z}})^{\kappa }{\frac {1}{(2i)^{\lambda }}}(z-{\bar {z}})^{\lambda }\\&=\displaystyle \sum _{\kappa ,\lambda }{\frac {a_{\kappa \lambda }}{2^{\kappa +\lambda }i^{\lambda }}}\sum _{\alpha =0}^{\kappa }{\binom {\kappa }{\alpha }}z^{\alpha }{\bar {z}}^{\kappa -\alpha }\cdot \sum _{\beta =0}^{\lambda }(-1)^{\lambda -\beta }{\binom {\lambda }{\beta }}z^{\beta }{\bar {z}}^{\lambda -\beta }\\&=\displaystyle \sum _{\kappa ,\lambda }\sum _{\alpha =0}^{\kappa }\sum _{\beta =0}^{\lambda }{\frac {a_{\kappa \lambda }}{2^{\kappa +\lambda }i^{\lambda }}}{\binom {\kappa }{\alpha }}(-1)^{\lambda -\beta }{\binom {\lambda }{\beta }}z^{\alpha +\beta }{\bar {z}}^{\kappa +\lambda -(\alpha +\beta )}\end{array}}}
Nun führen wir eine Indextransformation durch, wir ersetzen die Summation über
α
{\displaystyle \alpha }
durch eine Summation über
μ
:=
α
+
β
{\displaystyle \mu :=\alpha +\beta }
. Wegen
0
≤
α
≤
κ
,
0
≤
β
≤
λ
{\displaystyle 0\leq \alpha \leq \kappa ,0\leq \beta \leq \lambda }
läuft
μ
{\displaystyle \mu }
von
0
{\displaystyle 0}
bis
κ
+
λ
{\displaystyle \kappa +\lambda }
. Für festes
μ
{\displaystyle \mu }
müssen wir für
β
{\displaystyle \beta }
die Werte zwischen
0
{\displaystyle 0}
und
λ
{\displaystyle \lambda }
betrachten, für die
0
≤
μ
−
β
≤
κ
{\displaystyle 0\leq \mu -\beta \leq \kappa }
, also
μ
−
κ
≤
β
≤
μ
{\displaystyle \mu -\kappa \leq \beta \leq \mu }
ist. Daher läuft
β
{\displaystyle \beta }
von
max
(
0
,
μ
−
κ
)
{\displaystyle \max(0,\mu -\kappa )}
bis
min
(
μ
,
λ
)
{\displaystyle \min(\mu ,\lambda )}
. Wir erhalten
p
(
z
)
=
∑
κ
,
λ
∑
α
=
0
κ
∑
β
=
0
λ
a
κ
λ
2
κ
+
λ
i
λ
(
κ
α
)
(
−
1
)
λ
−
β
(
λ
β
)
z
α
+
β
z
¯
κ
+
λ
−
(
α
+
β
)
=
∑
κ
,
λ
∑
μ
=
0
κ
+
λ
∑
β
=
max
(
0
,
μ
−
κ
)
min
(
μ
,
λ
)
a
κ
λ
2
κ
+
λ
i
λ
(
κ
μ
−
β
)
(
−
1
)
λ
−
β
(
λ
β
)
z
μ
z
¯
κ
+
λ
−
μ
{\displaystyle {\begin{array}{rl}p(z)&=\displaystyle \sum _{\kappa ,\lambda }\sum _{\alpha =0}^{\kappa }\sum _{\beta =0}^{\lambda }{\frac {a_{\kappa \lambda }}{2^{\kappa +\lambda }i^{\lambda }}}{\binom {\kappa }{\alpha }}(-1)^{\lambda -\beta }{\binom {\lambda }{\beta }}z^{\alpha +\beta }{\bar {z}}^{\kappa +\lambda -(\alpha +\beta )}\\&=\displaystyle \sum _{\kappa ,\lambda }\sum _{\mu =0}^{\kappa +\lambda }\sum _{\beta =\max(0,\mu -\kappa )}^{\min(\mu ,\lambda )}{\frac {a_{\kappa \lambda }}{2^{\kappa +\lambda }i^{\lambda }}}{\binom {\kappa }{\mu -\beta }}(-1)^{\lambda -\beta }{\binom {\lambda }{\beta }}z^{\mu }{\bar {z}}^{\kappa +\lambda -\mu }\end{array}}}
Nun tauschen wir die Summationsreihenfolge, es ist
0
≤
μ
≤
k
+
ℓ
{\displaystyle 0\leq \mu \leq k+\ell }
und für festes
μ
{\displaystyle \mu }
muss stets
κ
+
λ
≥
μ
{\displaystyle \kappa +\lambda \geq \mu }
gelten, d. h.
λ
≥
μ
−
κ
{\displaystyle \lambda \geq \mu -\kappa }
,
wir erhalten
p
(
z
)
=
∑
κ
,
λ
∑
μ
=
0
κ
+
λ
∑
β
=
max
(
0
,
μ
−
κ
)
min
(
μ
,
λ
)
a
κ
λ
2
κ
+
λ
i
λ
(
κ
μ
−
β
)
(
−
1
)
λ
−
β
(
λ
β
)
z
μ
z
¯
κ
+
λ
−
μ
=
∑
μ
=
0
k
+
ℓ
∑
κ
=
0
μ
∑
λ
=
μ
−
κ
ℓ
∑
β
=
max
(
0
,
μ
−
κ
)
min
(
μ
,
λ
)
a
κ
λ
2
κ
+
λ
i
λ
(
κ
μ
−
β
)
(
−
1
)
λ
−
β
(
λ
β
)
z
μ
z
¯
κ
+
λ
−
μ
{\displaystyle {\begin{array}{rl}p(z)&=\displaystyle \sum _{\kappa ,\lambda }\sum _{\mu =0}^{\kappa +\lambda }\sum _{\beta =\max(0,\mu -\kappa )}^{\min(\mu ,\lambda )}{\frac {a_{\kappa \lambda }}{2^{\kappa +\lambda }i^{\lambda }}}{\binom {\kappa }{\mu -\beta }}(-1)^{\lambda -\beta }{\binom {\lambda }{\beta }}z^{\mu }{\bar {z}}^{\kappa +\lambda -\mu }\\&=\displaystyle \sum _{\mu =0}^{k+\ell }\sum _{\kappa =0}^{\mu }\sum _{\lambda =\mu -\kappa }^{\ell }\sum _{\beta =\max(0,\mu -\kappa )}^{\min(\mu ,\lambda )}{\frac {a_{\kappa \lambda }}{2^{\kappa +\lambda }i^{\lambda }}}{\binom {\kappa }{\mu -\beta }}(-1)^{\lambda -\beta }{\binom {\lambda }{\beta }}z^{\mu }{\bar {z}}^{\kappa +\lambda -\mu }\end{array}}}
Nun noch eine letzte Indextransformation. Wir ersetzen
κ
{\displaystyle \kappa }
durch
ν
:=
κ
+
λ
−
μ
{\displaystyle \nu :=\kappa +\lambda -\mu }
.
ν
{\displaystyle \nu }
läuft von
0
{\displaystyle 0}
bis
k
+
ℓ
{\displaystyle k+\ell }
. Für festes
ν
{\displaystyle \nu }
ist
κ
=
ν
+
μ
−
λ
{\displaystyle \kappa =\nu +\mu -\lambda }
. Es sind also nur die Werte von
λ
{\displaystyle \lambda }
zulässig, für die
0
≤
ν
+
μ
−
λ
≤
μ
{\displaystyle 0\leq \nu +\mu -\lambda \leq \mu }
ist, d. h.
ν
≤
λ
≤
ν
+
μ
{\displaystyle \nu \leq \lambda \leq \nu +\mu }
gilt. Wir erhalten
p
(
z
)
=
∑
μ
=
0
k
+
ℓ
∑
κ
=
0
μ
∑
λ
=
μ
−
κ
ℓ
∑
β
=
max
(
0
,
μ
−
κ
)
min
(
μ
,
λ
)
a
κ
λ
2
κ
+
λ
i
λ
(
κ
μ
−
β
)
(
−
1
)
λ
−
β
(
λ
β
)
z
μ
z
¯
κ
+
λ
−
μ
=
∑
μ
=
0
k
+
ℓ
∑
ν
=
0
k
+
ℓ
(
∑
λ
=
ν
min
(
ℓ
,
ν
+
μ
)
∑
β
=
λ
−
ν
min
(
μ
,
λ
)
a
μ
+
ν
−
λ
,
λ
2
μ
+
ν
i
λ
(
μ
+
ν
−
λ
μ
−
β
)
(
−
1
)
λ
−
β
(
λ
β
)
)
z
μ
z
¯
ν
{\displaystyle {\begin{array}{rl}p(z)&=\displaystyle \sum _{\mu =0}^{k+\ell }\sum _{\kappa =0}^{\mu }\sum _{\lambda =\mu -\kappa }^{\ell }\sum _{\beta =\max(0,\mu -\kappa )}^{\min(\mu ,\lambda )}{\frac {a_{\kappa \lambda }}{2^{\kappa +\lambda }i^{\lambda }}}{\binom {\kappa }{\mu -\beta }}(-1)^{\lambda -\beta }{\binom {\lambda }{\beta }}z^{\mu }{\bar {z}}^{\kappa +\lambda -\mu }\\&=\displaystyle \sum _{\mu =0}^{k+\ell }\sum _{\nu =0}^{k+\ell }\left(\sum _{\lambda =\nu }^{\min(\ell ,\nu +\mu )}\sum _{\beta =\lambda -\nu }^{\min(\mu ,\lambda )}{\frac {a_{\mu +\nu -\lambda ,\lambda }}{2^{\mu +\nu }i^{\lambda }}}{\binom {\mu +\nu -\lambda }{\mu -\beta }}(-1)^{\lambda -\beta }{\binom {\lambda }{\beta }}\right)z^{\mu }{\bar {z}}^{\nu }\end{array}}}
Also ist
b
μ
ν
=
∑
λ
=
ν
min
(
ℓ
,
ν
+
μ
)
∑
β
=
λ
−
ν
min
(
μ
,
λ
)
a
μ
+
ν
−
λ
,
λ
2
μ
+
ν
i
λ
(
μ
+
ν
−
λ
μ
−
β
)
(
−
1
)
λ
−
β
(
λ
β
)
,
0
≤
μ
,
ν
≤
k
+
ℓ
{\displaystyle b_{\mu \nu }=\sum _{\lambda =\nu }^{\min(\ell ,\nu +\mu )}\sum _{\beta =\lambda -\nu }^{\min(\mu ,\lambda )}{\frac {a_{\mu +\nu -\lambda ,\lambda }}{2^{\mu +\nu }i^{\lambda }}}{\binom {\mu +\nu -\lambda }{\mu -\beta }}(-1)^{\lambda -\beta }{\binom {\lambda }{\beta }},\qquad 0\leq \mu ,\nu \leq k+\ell }