-
Notifications
You must be signed in to change notification settings - Fork 1
/
03_differentialrechnung.tex
571 lines (470 loc) · 21.7 KB
/
03_differentialrechnung.tex
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
\chapter{%
Zur Dif"|ferentialrechnung von Funktionen einer Variablen%
}
\section{%
Die Definition der Ableitung%
}
\enlargethispage{10mm}
Sei $f: X \subset \mathbb{K} \rightarrow \mathbb{K}^n$ eine Funktion
mit $X$ of"|fen, d.\,h. für einen Punkt $x_0 \in X$ ist \\
$\exists_{\varepsilon > 0}\; U_\varepsilon(x_0) \subset X$.
Daraus folgt $x_0 + h \in X$ für $|h| < \varepsilon$. \\
$\varphi(h, x_0) =$ {\large $\frac{f(x_0 + h) - f(x_0)}{h}$} heißt
\textbf{Dif"|ferenzenquotient} ($|h| < \varepsilon$, $h \not= 0$).
$f$ heißt im Punkt $x_0 \in X$ \textbf{dif"|ferenzierbar}, falls der Grenzwert \\
$\lim_{h \to 0} \varphi(h, x_0) =: f'(x_0) = f'|_{x=x_0} =$
{\large $\frac{df}{dx}$}$\big|_{x=x_0}$ existiert. \\
$f$ heißt dif"|ferenzierbar in $X$, falls $f$ in allen Punkten $x_0 \in X$
dif"|ferenzierbar ist.
\linie
Für Funktionen $f: X \subset \mathbb{C} \rightarrow \mathbb{C}$ kann
man für $x_0 \in X \cap \mathbb{R}$ die \textbf{komplexe bzw. reelle Ableitung}
$(\mathbb{C}) - f'(x_0) = \lim_{h \to 0,\; h \in \mathbb{C}}
\frac{f(x_0 + h) - f(x_0)}{h}$ bzw.
$(\mathbb{R}) - f'(x_0) = \lim_{h \to 0,\; h \in \mathbb{R}}
\frac{f|_{\mathbb{R}}(x_0 + h) - f|_{\mathbb{R}}(x_0)}{h}$
betrachten.
Existieren die Grenzwerte, so heißt $f$ \textbf{komplex bzw. reell
dif"|ferenzierbar}.
\textbf{Satz}: Ist $f: X \subset \mathbb{C} \rightarrow \mathbb{C}$ in
$x_0 \in \mathbb{R} \cap X$ $(\mathbb{C})$-dif"|ferenzierbar, so ist sie auch
$(\mathbb{R})$-dif"|ferenzierbar und
$(\mathbb{C}) - f'(x_0) = (\mathbb{R}) - f'(x_0)$.
\emph{Die Umkehrung gilt nicht!}
\textbf{Satz}: Eine komplexwertige Funktion
$f: X \subset \mathbb{R} \rightarrow \mathbb{C}$, $f = g + ik$
($g,k: X \subset \mathbb{R} \rightarrow \mathbb{R}$) ist genau dann
reell dif"|ferenzierbar, wenn Real- und Imaginärteil reell dif"|ferenzierbar sind.
\textbf{Satz}: Ist $f: X \subset \mathbb{C} \rightarrow \mathbb{R}$
in $z_0 \in X$ komplex dif"|ferenzierbar, so ist
$(\mathbb{C}) - f'(z_0) = 0$.
\section{%
Die \textsc{Landau}-Symbole%
}
Seien $M$ ein metrischer Raum,
$f, g: X \subset M \rightarrow \mathbb{K}^n$
sowie $x_0 \in \acc(X)$.
\textbf{\textsc{Landau}-Symbole}:
$f \overset{x \to x_0}{=} \mathcal{O}(g) \;\Leftrightarrow\;
\exists_{C \in \mathbb{R}} \exists_{\delta > 0}
\forall_{x \in X \cap U_\delta(x_0)}\;
\Vert f(x) \Vert \le C \Vert g(x) \Vert$, \\
$f \overset{x \to x_0}{=} o(g) \;\Leftrightarrow\;
\forall_{\varepsilon > 0} \exists_{\delta = \delta(\varepsilon)}
\forall_{x \in X \cap U_\delta(x_0)}\;
\Vert f(x) \Vert \le \varepsilon \Vert g(x) \Vert$
\qquad (in $\mathbb{K}$ ist die Norm der Betrag)
\linie
\textbf{Satz}: Sei $x \to x_0 \in \acc(X)$. \qquad
Dann gilt $f = o(g) \;\Rightarrow\; f = \mathcal{O}(g)$, \\
$f_1 = \mathcal{O}(g) \land f_2 = \mathcal{O}(g) \;\Rightarrow\;
f_1 \pm f_2 = \mathcal{O}(g)$, \qquad
$f_1 = o(g) \land f_2 = o(g) \;\Rightarrow\; f_1 \pm f_2 = o(g)$ sowie
$f_1 = o(g) \land f_2 = \mathcal{O}(g) \;\Rightarrow\;
f_1 \pm f_2 = \mathcal{O}(g)$.
\textbf{Satz}: Seien $f, g: X \subset M \rightarrow \mathbb{K}^n$,
$\gamma, \psi: X \subset M \rightarrow \mathbb{K}$.
Dann gilt \\
$\psi = \mathcal{O}(\gamma) \land f = \mathcal{O}(g) \;\Rightarrow\;
\psi f = \mathcal{O}(\gamma g)$, \qquad
$\psi = o(\gamma) \land f = \mathcal{O}(g) \;\Rightarrow\;
\psi f = o(\gamma g)$ sowie \\
$\psi = \mathcal{O}(\gamma) \land f = o(g) \;\Rightarrow\;
\psi f = o(\gamma g)$.
\textbf{Schreibweise}:
$f_1 - f_2 = \mathcal{O}(g) \;\Leftrightarrow\; f_1 = f_2 + \mathcal{O}(g)$,
\qquad $f_1 - f_2 = o(g) \;\Leftrightarrow\; f_1 = f_2 + o(g)$
\textbf{Anmerkung}:
Ist $f: X \subset \mathbb{K} \rightarrow \mathbb{K}^n$ und $x \to x_0 = 0$,
dann ist $f(x) = o(x) \;\Leftrightarrow\; f(x) = x \tilde{f}(x)$
mit $\tilde{f}(x) = o(1)$ \quad
(bzw. $f(x) = \mathcal{O}(x) \;\Leftrightarrow\; f(x) = x \tilde{f}(x)$
mit $\tilde{f}(x) = \mathcal{O}(1)$).
\linie
\textbf{Anwendungen}:
\begin{itemize}
\item $f \overset{x \to x_0}{=} \mathcal{O}(1)$
$\;\Leftrightarrow\;$ $f$ ist in einer geeigneten $\delta$-Umgebung
von $x_0$ beschränkt
\item $f \overset{x \to x_0}{=} o(1)$
$\;\Leftrightarrow\;$ $(\lim_{x \to x_0} f(x) = 0) \land
(x_0 \in X \Rightarrow f(x_0) = 0)$
\item $f(x_0 + h) \overset{h \to 0}{=} f(x_0) + o(1)$
$\;\Leftrightarrow\;$ $f$ ist stetig in $x_0$
\item $f(x_0 + h) - f(x_0) \overset{h \to 0}{=} hF + o(h)$
$\;\Leftrightarrow\;$ $f$ ist in $x_0$ dif"|ferenzierbar und $f'(x_0) = F$
\end{itemize}
\textbf{Folgerung}:
Ist $f$ im Punkt $x_0$ dif"|ferenzierbar, so ist $f$ im Punkt $x_0$ stetig. \\
Die Umkehrung gilt nicht!
\section{%
Das Rechnen mit Ableitungen%
}
Seien $\mathbb{K} \in \{\mathbb{R}, \mathbb{C}\}$, $X \subset \mathbb{K}$
of"|fen, $x_0 \in X$, \quad
$f, f_1, f_2: X \subset \mathbb{K} \rightarrow \mathbb{K}^n$,
$g: X \subset \mathbb{K} \rightarrow \mathbb{K}$, \\
$f, f_1, f_2, g$ im Punkt $x_0 \in X$ dif"|ferenzierbar, \\
$\psi: Y \subset \mathbb{K} \rightarrow \mathbb{K}$, $Y$ of"|fen, $y_0 \in Y$
mit $\psi(y_0) = x_0$, $\psi$ im Punkt $y_0 \in Y$ dif"|ferenzierbar.
Dann ist $(f_1 + f_2)'|_{x=x_0} = f_1'|_{x=x_0} + f_2'|_{x=x_0}$, \quad
$(\alpha f)'|_{x=x_0} = \alpha (f'|_{x=x_0})$, \\
$(gf)'|_{x=x_0} = g'|_{x=x_0} f(x_0) + g(x_0) f'|_{x=x_0}$ \quad sowie \quad
$(f \circ \psi)'|_{y=y_0} = f'|_{x=x_0=\psi(y_0)} \cdot \psi'|_{y=y_0}$.
\textbf{Folgerung}: Seien $X \subset \mathbb{K}$ of"|fen, $x_0 \in X$,
$f, g: X \subset \mathbb{K} \rightarrow \mathbb{K}$, $g(x) \not= 0$ für
alle $x \in X$, \\
$f, g$ dif"|ferenzierbar in $x_0 \in X$. \quad
Dann ist {\large $\left(\frac{f}{g}\right)'\Big|_{x=x_0} =
\frac{f'(x_0) g(x_0) - f(x_0) g'(x_0)}{(g(x_0))^2}$}.
\textbf{Satz}:
Seien $X, Y \subset \mathbb{K}$ of"|fen, $x_0 \in X$, $y_0 \in Y$,
$f: X \rightarrow Y$ bijektiv mit $y_0 = f(x_0)$, \\
$f^{-1}$ stetig im Punkt $y_0$ sowie $f$ dif"|ferenzierbar in $x_0$ mit
$f'(x_0) \not= 0$. \\
Dann ist $f^{-1}$ in $y_0$ dif"|ferenzierbar mit
$(f^{-1})'(y_0) =$ {\large $\frac{1}{f'(x_0)}$}.
\section{%
Ableitungen wichtiger Funktionen%
}
\begin{tabular}{lllll}
$(\text{const.})' = 0$ &
$(z)' = 1$ &
$(z^\alpha)' = \alpha z^{\alpha - 1}$ \\
$(e^z)' = e^z$ &
$(\Ln z)' =$ {\large $\frac{1}{z}$} \\ \hline
$(\sin z)' = \cos z$ &
$(\cos z)' = -\sin z$ &
$(\tan z)' =$ {\large $\frac{1}{\cos^2 z}$} &
$(\cot z)' =$ {\large $-\frac{1}{\sin^2 z}$} \\
$(\sinh z)' = \cosh z$ &
$(\cosh z)' = \sinh z$ &
$(\tanh z)' =$ {\large $\frac{1}{\cosh^2 z}$} &
$(\coth z)' =$ {\large $-\frac{1}{\sinh^2 z}$} \\ \hline
$(\arcsin z)' =$ {\large $\frac{1}{\sqrt{1 - z^2}}$} &
$(\arccos z)' =$ {\large $-\frac{1}{\sqrt{1 - z^2}}$} &
$(\arctan z)' =$ {\large $\frac{1}{1 + z^2}$} &
$(\arccot z)' =$ {\large $-\frac{1}{1 + z^2}$} \\
$(\arsinh z)' =$ {\large $\frac{1}{\sqrt{z^2 + 1}}$} &
$(\arcosh z)' =$ {\large $\frac{1}{\sqrt{z^2 - 1}}$} &
$(\artanh z)' =$ {\large $\frac{1}{1 - z^2}$} &
$(\arcoth z)' =$ {\large $\frac{1}{1 - z^2}$}
\end{tabular}
\section{%
Die Sätze von \textsc{Fermat}, \textsc{Rolle}, \textsc{Cauchy} und
\textsc{Lagrange}%
}
Wir betrachten nun reelle Ableitungen: $f: [a,b] \rightarrow \mathbb{R}$,
$a < b$.
\textbf{Satz von \textsc{Fermat}}:
Sei $f \in C([a,b])$,
$c \in \left]a,b\right[$ mit $f$ in $c$ dif"|fb. sowie \\
$f(c) = \max_{x \in [a,b]} f(x)$ bzw. $f(c) = \min_{x \in [a,b]} f(x)$. \qquad
Dann ist $f'(c) = 0$.
\textbf{Satz von \textsc{Rolle}}:
Sei $f \in C([a,b])$, $f$ in $\left]a,b\right[$ dif"|fb. sowie
$f(a) = f(b)$. \\
Dann gibt es ein $c \in \left]a,b\right[$, sodass $f'(c) = 0$.
\textbf{Satz von \textsc{Cauchy}}:
Seien $f,g \in C([a,b])$, $f,g$ in $\left]a,b\right[$ dif"|fb. sowie
$g'(x) \not= 0$ für alle $x \in \left]a,b\right[$. \\
Dann gibt es ein $c \in \left]a,b\right[$, sodass
{\large $\frac{f(b) - f(a)}{g(b) - g(a)} = \frac{f'(c)}{g'(c)}$}.
\textbf{Satz von \textsc{Lagrange}}:
Sei $f \in C([a,b])$ in $\left]a,b\right[$ dif"|fb. \\
Dann gibt es ein $c \in \left]a,b\right[$, sodass
$f(b) - f(a) = (b - a) \cdot f'(c)$.
\section{%
Hauptsatz der Dif"|ferentialrechnung%
}
Sei $f: X \subset \mathbb{K} \rightarrow \mathbb{K}^n$, mit $X$ of"|fen und
$\overline{ab} \subset X$, wobei $\overline{ab}$ für $a, b \in X$ definiert ist
als $\overline{ab} = \{x \in \mathbb{K} \;|\;
x = a +$ {\large $\frac{b - a}{|b - a|}$} $t,\; t \in [0, |b - a|]\}$ und
{\scriptsize $\overset{\circ}{\overline{ab}}$}
$= \overline{ab} \setminus \{a, b\}$.
\textbf{Hauptsatz der Dif"|ferentialrechnung}:
Sei $f \circ \psi$ stetig auf $[0, |b - a|]$ und dif"|ferenzierbar für
$t \in \left]0, |b - a|\right[$ (d.\,h. $f$ stetig auf $\overline{ab}$ und
dif"|ferenzierbar auf {\scriptsize $\overset{\circ}{\overline{ab}}$}), wobei
$\psi(t) = a +$ {\large $\frac{b - a}{|b - a|}$} $t$. \\
Dann ist $\Vert f(b) - f(a) \Vert \le
\sup_{x \in \overset{\circ}{\overline{ab}}} \Vert f'(x) \Vert \cdot |b - a|$.
\pagebreak
\section{%
Ableitungen höherer Ordnung%
}
Sei $f: X \subset \mathbb{K} \rightarrow \mathbb{K}^n$ mit $X$ of"|fen.
Ist diese Funktion in einer $\varepsilon$-Umgebung von $x_0 \in X$
mit $U_\varepsilon(x_0) \subset X$ dif"|fb., so kann die Ableitung
als Funktion $f': U_\varepsilon(x_0) \rightarrow \mathbb{K}^n$ dargestellt
werden.
\textbf{höhere Ableitungen}:
Ist $f': U_\varepsilon(x_0) \rightarrow \mathbb{K}^n$ im Punkt $x_0$
dif"|ferenzierbar, so heißt
$(f')'(x_0) =:$ {\large $\frac{d^2 f}{dx^2}$}$\big|_{x=x_0} = f''(x_0) =
f^{(2)}(x_0)$ die \textbf{zweite Ableitung von $f$}. \\
Die Definition kann iterativ fortgesetzt werden:
Ist $f^{(m-1)}: U_\varepsilon(x_0) \rightarrow \mathbb{K}^n$ in $x_0$
dif"|ferenzierbar, so ist analog
$(f^{(m-1)})'(x_0) =:$ {\large $\frac{d^m f}{dx^m}$}$\big|_{x=x_0} =
f^{(m)}(x_0)$ die \textbf{$m$-te Ableitung von $f$}.
\textbf{Schreibweise}: \\
$C^m(X, \mathbb{K}^n) = \{f: X \subset \mathbb{K} \rightarrow \mathbb{K}^n
\;|\; f \text{ auf } X \text{ } m \text{-fach dif"|ferenzierbar},\;
f^{(m)} \text{ auf } X \text{ stetig}\}$, \\
$C^\infty(X, \mathbb{K}^n) = \{f: X \subset \mathbb{K} \rightarrow \mathbb{K}^n
\;|\; f \text{ beliebig oft auf } X \text{ dif"|ferenzierbar}\}$
\textbf{Satz von \textsc{Leibniz}}:
Seien $f: X \subset \mathbb{K} \rightarrow \mathbb{K}^n$ und
$g: X \subset \mathbb{K} \rightarrow \mathbb{K}$ ($X$ of"|fen)
$m$-fach dif"|fb. in $X$. \\
Dann ist auch $(g \cdot f)$ $m$-fach dif"|ferenzierbar und
$(gf)^{(m)}(x_0) = \sum_{k=0}^m \binom{m}{k} g^{(k)}(x_0) f^{(m-k)}(x_0)$
(dabei sei $g^{(0)} = g$ und $f^{(0)} = f$).
\section{%
Der Satz von \textsc{Taylor}%
}
Sei $f: X \subset \mathbb{K} \rightarrow \mathbb{K}^n$ ($X$ of"|fen)
in $x_0 \in X$ $m$-fach dif"|ferenzierbar. \\
Dann ist $f(x_0 + h) = f(x_0) + \sum_{k=1}^m \frac{1}{k!} f^{(k)}(x_0) h^k +
r_m(h)$ mit $r_m(h) = o(h^m)$ für $h \to 0$.
\section{%
Monotonie und Extremwerte von Funktionen%
}
\textbf{Satz}:
Sei $f: [a,b] \rightarrow \mathbb{R}^n$ stetig auf $[a,b]$ und dif"|ferenzierbar
in $\left]a,b\right[$. \\
Dann ist $f$ konstant auf $[a,b]$ genau dann, wenn $f'(x) = 0$ für alle
$x \in \left]a,b\right[$ ist.
\textbf{Folgerung}:
Seien $f, g: [a,b] \rightarrow \mathbb{R}^n$ stetig auf $[a,b]$
und dif"|ferenzierbar in $\left]a,b\right[$. \\
Dann folgt aus $f'(x) = g'(x)$ für alle $x \in \left]a,b\right[$, dass
$f(x) = g(x) + \text{const.}$ ist.
\textbf{Monotonie von Funktionen}:
Sei $f: [a,b] \rightarrow \mathbb{R}$. \\
$f\!\!\uparrow \quad\Leftrightarrow\quad
(x_1 < x_2 \;\Rightarrow\; f(x_1) \le f(x_2))$, \qquad
$f\!\!\upuparrows \quad\Leftrightarrow\quad
(x_1 < x_2 \;\Rightarrow\; f(x_1) < f(x_2))$
\textbf{Satz}:
Sei $f: [a,b] \rightarrow \mathbb{R}$ stetig auf $[a,b]$ sowie
dif"|ferenzierbar in $\left]a,b\right[$. \\
Dann ist $f\!\!\uparrow \quad\Leftrightarrow\quad
\forall_{x \in \left]a,b\right[}\; f'(x) \ge 0$ \quad sowie \\
$f\!\!\upuparrows \quad\Leftrightarrow\quad
(\forall_{x \in \left]a,b\right[}\; f'(x) \ge 0) \land
\lnot(\exists_{\alpha, \beta \in \left]a,b\right[,\; \alpha < \beta}
\forall_{x \in [\alpha, \beta]}\; f'(x) = 0)$.
\linie
\textbf{globale Extremwerte}:
$f: X \subset \mathbb{R} \rightarrow \mathbb{R}$ nimmt im Punkt $c \in X$
ein globales Maximum (bzw. Minimum) an, falls
$f(c) \ge f(x)$ (bzw. $f(c) \le f(x)$) für alle $x \in X$.
\textbf{notwendige Bedingung (globale Extrema)} (Satz von \textsc{Fermat}):
Seien $f: [a,b] \rightarrow \mathbb{R}$ stetig, in $\left]a,b\right[$ dif"|fb.
und $c \in \left]a,b\right[$ mit $f(c) = \max_{x \in [a,b]} f(x)$. \qquad
Dann ist $f'(c) = 0$.
\textbf{hinreichende Bedingung (globale Extrema)}:
Seien $f: [a,b] \rightarrow \mathbb{R}$ stetig, in $\left]a,b\right[$ dif"|fb.
und $c \in \left]a,b\right[$ mit $f'(c) = 0$, wobei
$f'(x) \ge 0$ für $x < c$ und $f'(x) \le 0$ für $x > c$
($x \in \left]a,b\right[$). \\
Dann ist $f(c) = \max_{x \in [a,b]} f(x)$.
\textbf{Folgerung (doppelte Ableitung)}:
Seien $f: [a,b] \rightarrow \mathbb{R}$ stetig, in $\left]a,b\right[$ 2-fach
dif"|fb. und $c \in \left]a,b\right[$ mit $f'(c) = 0$ sowie $f''(x) \le 0$ für
alle $x \in \left]a,b\right[$. \qquad
Dann ist $f(c) = \max_{x \in [a,b]} f(x)$.
\linie
\pagebreak
\textbf{lokale Extremwerte}:
$f: X \subset \mathbb{R} \rightarrow \mathbb{R}$ nimmt im Punkt $c \in X$
ein lokales Maximum (bzw. Minimum) an, falls
$\exists_{\varepsilon > 0} \forall_{x \in X \cap U_\varepsilon(c)}\;
f(c) \ge f(x)$ (bzw. $f(c) \le f(x)$).
\textbf{notwendige Bedingung (lokale Extrema)}:
Sei $f: [a,b] \rightarrow \mathbb{R}$ stetig, in $\left]a,b\right[$ dif"|fb.
und $c \in \left]a,b\right[$, wobei $f$ in $c$ einen lokalen Extremwert
annimmt. \qquad
Dann ist $f'(c) = 0$.
\textbf{hinreichende Bedingung (lokale Extrema)}:
Sei $f: [a,b] \rightarrow \mathbb{R}$ stetig, in $\left]a,b\right[$ dif"|fb.
sowie in $c \in \left]a,b\right[$ 2-fach dif"|fb., wobei
$f'(c) = 0$ und $f''(c) < 0$. \\
Dann nimmt $f$ in $c$ ein lokales Maximum an.
\textbf{$n$-fache Ableitung (Extrema)}:
Sei $f: [a,b] \rightarrow \mathbb{R}$ in $\left]a,b\right[$ $n-1$-fach dif"|fb.
sowie in $c \in \left]a,b\right[$ $n$-fach dif"|fb., wobei
$f'(c) = \cdots = f^{(n-1)}(c) = 0$ und $f^{(n)} \not= 0$. \\
Dann ist, falls $n$ gerade ist, $c$ ein lokales Maximum falls $f^{(n)}(c) < 0$
bzw. ein lokales Minimum falls $f^{(n)}(c) > 0$. \qquad
Ist $n$ ungerade, so ist $c$ kein lokaler Extremwert.
\section{%
Konvexe und konkave Funktionen%
}
Sei $f: [a,b] \rightarrow \mathbb{R}$.
\textbf{konvexe und konkave Funktionen}:
$f$ heißt konvex \\
$\;\Leftrightarrow\; \forall_{x_1, x_2 \in [a,b],\; x_1 < x_2}
\forall_{t \in [0,1]}\; f(tx_1 + (1 - t)x_2) \le t f(x_1) + (1 - t) f(x_2)$. \\
$f$ heißt konkav $\;\Leftrightarrow\;$ $-f$ ist konvex.
\textbf{Äquivalente Definition (Ableitung)}:
Sei $f$ stetig auf $[a,b]$ und dif"|ferenzierbar in $\left]a,b\right[$. \\
Dann ist $f$ konvex $\;\Leftrightarrow\; f'\!\!\uparrow$ \qquad und \qquad
$f$ konkav $\;\Leftrightarrow\; f'\!\!\downarrow$.
\textbf{doppelte Ableitung}:
Sei $f$ stetig auf $[a,b]$, 2-fach dif"|fb. in $\left]a,b\right[$ sowie
$f''(x) \ge 0$ für alle $x \in \left]a,b\right[$. \qquad
Dann ist $f$ konvex.
\linie
\textbf{Wendepunkt}:
Sei $f$ in $\left]a,b\right[$ dif"|ferenzierbar. \\
$c \in \left]a,b\right[$ heißt Wendepunkt, falls $f'(c)$ ein lokales Extremum
ist.
\textbf{notwendige Bedingung (Wendepunkte)}:
Seien $f$ in $\left]a,b\right[$ 2-fach dif"|fb. und $c \in \left]a,b\right[$
ein Wendepunkt. \qquad
Dann ist $f''(c) = 0$.
\textbf{$n$-fache Ableitung (Wendepunkte)}:
Sei $f$ in $\left]a,b\right[$ $n$-fach dif"|fb. sowie in
$c \in \left]a,b\right[$ $n+1$-fach dif"|fb., wobei
$f^{(2)}(c) = \cdots = f^{(n)}(c) = 0$ und $f^{(n+1)}(c) \not= 0$. \\
Dann ist $c$ ein Wendepunkt, falls $n$ gerade, und kein Wendepunkt, falls
$c$ ungerade ist.
\pagebreak
\section{%
\texorpdfstring
{Das Auf"|lösen von Unbestimmtheiten vom Typ $0/0$ und $\infty/\infty$}%
{Das Auf"|lösen von Unbestimmtheiten vom Typ 0/0 und ∞/∞}%
}
\textbf{Typ $0/0$}:
Seien $f, g: [a,b] \rightarrow \mathbb{R}$ ($\mathbb{C}$, $\mathbb{R}^n$,
$\mathbb{C}^n$) und $x_0 \in \left]a,b\right[$ mit $f, g$ in $x_0$ dif"|fb., \\
$f(x_0) = g(x_0) = 0$
sowie $g'(x_0) \not= 0$. \qquad
Dann existiert der Grenzwert $\lim_{x \to x_0}$
{\large $\frac{f(x)}{g(x)} = \frac{f'(x_0)}{g'(x_0)}$}.
\textbf{Verallgemeinerung}:
Seien $f, g: [a,b] \rightarrow \mathbb{R}$ ($\mathbb{C}$) und
$x_0 \in \left]a,b\right[$ mit $f(x_0) = g(x_0) = 0$,
$f'(x_0) = g'(x_0) = 0$, \dots, $f^{(n-1)}(x) = g^{(n-1)}(x) = 0$,
$\exists f^{(n)}(x_0)$, $\exists g^{(n)}(x_0)$, wobei $g^{(n)}(x_0) \not= 0$.
Dann existiert der Grenzwert $\lim_{x \to x_0}$
{\large $\frac{f(x)}{g(x)} = \frac{f^{(n)}(x_0)}{g^{(n)}(x_0)}$}.
\linie
\textbf{Regel von \textsc{Bernoulli} und \textsc{l'Hôspital}}:
Seien $f, g: \left]a,b\right[ \rightarrow \mathbb{R}$ in $\left]a,b\right[$
dif"|fb., \\
$\lim_{x \to a} f(x) = \lim_{x \to a} g(x) = 0$ und $g'(x) \not= 0$
für $x \in \left]a,b\right[$.
Außerdem existiere der Grenzwert $\lim_{x \to a}$%
{\large $\frac{f'(x)}{g'(x)}$} $=: A$. \qquad
Dann existiert der Grenzwert $\lim_{x \to a}$%
{\large $\frac{f(x)}{g(x)}$} $= A$.
Dieser Satz gilt nur für reellwertige (nicht für komplexwertige) Funktionen!
Anwendung: bei Funktionen
$f, g: \left[b,+\infty\right[ \rightarrow \mathbb{R}$, $b > 0$,
wobei \\
$\lim_{x \to +\infty} f(x) = \lim_{x \to +\infty} g(x) = 0$ und
$A = \lim_{x \to +\infty} \frac{f'(x)}{g'(x)}$.
Dann ist $\lim_{x \to +\infty} \frac{f(x)}{g(x)} = A$. \\
(\emph{Variablentransformation} mit $x = \frac{1}{t}$)
\linie
\textbf{Typ $\infty/\infty$}:
Seien $f, g: \left]a,b\right[ \rightarrow \mathbb{R}$ in $\left]a,b\right[$
dif"|fb., $\lim_{x \to a} f(x) = \infty$, $\lim_{x \to a} g(x) = \infty$
und es existiere der Grenzwert $\lim_{x \to a}$%
{\large $\frac{f'(x)}{g'(x)}$} $=: A$. \qquad
Dann existiert der Grenzwert $\lim_{x \to a}$%
{\large $\frac{f(x)}{g(x)}$} $= A$.
Grenzwerte $f(x) \cdot g(x)$ vom Typ $\mathrel{\widehat{=}} 0 \cdot \infty
\mathrel{\widehat{=}}$ {\large $\frac{f(x)}{\frac{1}{g(x)}}$} kann man auf
$0/0$ zurückführen.
Grenzwerte $f(x)^{g(x)}$ mit $1^\infty$, $0^0$ oder $\infty^0$ kann man
mit $f(x)^{g(x)} = e^{g(x) \cdot \ln f(x)}$ auf $0 \cdot \infty$ zurückführen.
\section{%
Weitere Anwendungen der Dif"|ferentialrechnung%
}
\textbf{Tangente}: $y = f'(x_0) \cdot (x - x_0) + y_0$, \qquad
\textbf{Normale}: $y = -${\large $\frac{1}{f'(x_0)}$} $\cdot\; (x - x_0) + y_0$
\textbf{Dif"|ferentiation parametrisch gegebener Kurven}:
Gegeben seien die dif"|ferenzierbaren Funktionen
$\psi: \left]\alpha,\beta\right[ \rightarrow \left]a,b\right[$ sowie
$f: \left]a,b\right[ \rightarrow \mathbb{R}$.
Durch $x(t) = \psi(t)$ und $y(t) = f(\psi(t))$ sei für
$t \in \left]\alpha, \beta\right[$ eine Kurve gegeben.
Dann ist $f'(x_0) =$ {\large $\frac{\dot{y}(t_0)}{\dot{x}(t_0)}$}
für $x_0 = x(t_0)$.
\textbf{geradlinige Asymptote}:
$g(x) = ax + b$ ist eine (lokale) \emph{geradlinige Asymptote} von $f(x)$ für
$x \to +\infty$ (bzw. $x \to -\infty$), falls
$\lim_{x \to +\infty\; (\text{bzw. }-\infty)} (f(x) - g(x)) = 0$. \\
Dann ist $a = \lim_{x \to \pm\infty} \frac{f(x)}{x}$ und
$b = \lim_{x \to \pm\infty} (f(x) - ax)$.
\section{%
Der Satz von \textsc{Darboux}%
}
\textbf{Satz}:
Seien $f: \left]a,b\right[ \rightarrow \mathbb{R}$ dif"|fb. und
$x_1, x_2 \in \left]a,b\right[$ mit $x_1 < x_2$, wobei
$f'(x_1) \cdot f'(x_2) < 0$ ist. \\
Dann gibt es ein $x_0 \in \left]x_1,x_2\right[$, sodass $f'(x_0) = 0$.
\textbf{Satz von \textsc{Darboux}}:
Seien $f: \left]a,b\right[ \rightarrow \mathbb{R}$ dif"|fb. und
$x_1, x_2 \in \left]a,b\right[$ mit $x_1 < x_2$, wobei $f'(x_1) \not= f'(x_2)$.
Sei außerdem $\lambda \in \mathbb{R}$ mit $f'(x_1) < \lambda < f'(x_2)$ bzw.
$f'(x_2) < \lambda < f'(x_1)$. \\
Dann gibt es ein $x_0 \in \left]x_1,x_2\right[$, sodass $f'(x_0) = \lambda$.
\textbf{Satz}:
Sei $f: \left]a,b\right[ \rightarrow \mathbb{R}$ dif"|ferenzierbar. \quad
Dann besitzt $f'$ keine Unstetigkeit der ersten Art.
\section{%
Nullstellenberechnung%
}
Gegeben sei eine Funktion $f: [a,b] \rightarrow \mathbb{R}$ stetig mit
$f(a) f(b) < 0$, $f$ zweimal stetig dif"|fb. und $f'(x) \not= 0$,
$f''(x) \not= 0$ für alle $x \in \left]a,b\right[$ (d.\,h. $f', f''$ haben
konstantes Vorzeichen).
\textbf{Satz}: $\exists!_{\xi \in \left]a,b\right[}\; f(\xi) = 0$
\linie
\textbf{Regula falsi (Sehnenmethode)}:
Bei der \emph{Sehnenmethode} versucht man, $f$ durch die Sehne durch
$(a, f(a))$ und $(b, f(b))$ anzunähern.
Deren Gleichung lautet $g(x) = f(a) +$ {\large $\frac{f(b) - f(a)}{b - a}$}
$(x - a)$. \\
Für die Nullstelle $x_1 = a \;-$ {\large $\frac{b - a}{f(b) - f(a)}$}
$f(a) \in \left]a,b\right[$ gilt, dass $\xi \in \left]x_1,b\right[$
bzw. $\xi \in \left]a,x_1\right[$
(wenn $f', f''$ die gleichen bzw. unterschiedliche Vorzeichen haben).
Nun muss man nur noch in dem Intervall $[x_1,b]$
bzw. $[a,x_1]$ nach der Nullstelle $\xi$ suchen.
\textbf{Fehlerabschätzung}:
Sei $x_0 = a$, $x_n = x_{n-1} \;-$
{\large $\frac{b - x_{n-1}}{f(b) - f(x_{n-1})}$} $f(x_{n-1})$ bzw. \\
$x_0 = b$,
$x_n = x_{n-1} \;-$
{\large $\frac{x_{n-1} - a}{f(x_{n-1}) - f(a)}$} $f(x_{n-1})$. \\
Dann ist $\lim_{n \to \infty} x_n = \xi$, wobei
$|x_n - \xi| \le$ {\large $\frac{|f(x_n)|}{\min_{x \in [a,b]} |f'(x)|}$}.
\linie
\textbf{\textsc{Newton}-Verfahren (Tangentenmethode)}:
Beim \emph{Newton-Verfahren} versucht man, die Nullstelle $\xi$
durch Nullstellen der Ableitung zu bestimmen.
Für den Fall $\sgn(f') = \sgn(f'')$ gilt für die Tangentengleichung in
$x_0 = b$, dass $g(x) = f(b) + f'(b) \cdot (x - b)$, deren Nullstelle ist
$x_1 = b \;-$ {\large $\frac{f(b)}{f'(b)}$}.
Es gilt $x_1 \in [a,b]$.
Analog ist $x_1 = a \;-$ {\large $\frac{f(a)}{f'(a)}$} $\in [a,b]$ für
$\sgn(f') \not= \sgn(f'')$ (dann muss die Tangente in $x_0 = a$ bestimmt
werden).
Wiederum muss nun nur noch im Intervall $[a,x_1]$ bzw.
$[x_1,b]$ nach der Nullstelle $\xi$ gesucht werden.
\textbf{Fehlerabschätzung}:
Sei $x_0 = b$ bzw. $x_0 = a$ und
$x_n = x_{n-1} \;-$ {\large $\frac{f(x_{n-1})}{f'(x_{n-1})}$}. \\
Dann ist $\lim_{n \to \infty} x_n = \xi$, wobei
$\exists_{M > 0} \forall_{n \in \mathbb{N}}\;
|x_{n+1} - \xi| \le M |x_n - \xi|^2$.
\pagebreak