-
Notifications
You must be signed in to change notification settings - Fork 2
/
Copy pathaffine.tex
410 lines (325 loc) · 15.6 KB
/
affine.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
%\part*{Additional topics}
\addtocontents{toc}{\protect\contentsline{part}{\protect\numberline{}Additional topics}{}{}}
\chapter{Affine geometry}\label{chap:trans}
\section{Affine transformations}
\emph{Affine geometry} studies the so-called \index{incidence structure}\emph{incidence structure} of the Euclidean plane.
The incidence structure sees only which points lie on which lines and nothing else;
it does not directly see distances, angle measures, and many other things.
A bijection from the Euclidean plane to itself is called an \index{affine transformation}\emph{affine transformation} if it maps lines to lines;
that is, the image of any line is a line.
So, one can say that affine geometry studies the properties of the Euclidean plane that are preserved under affine transformations.
\begin{thm}{Exercise}\label{ex:affine-par}
Show that an affine transformation of the Euclidean plane sends any pair of parallel lines to a pair of parallel lines.
\end{thm}
The observation below follows since the lines are defined using the metric only.
\begin{thm}{Observation}
Any motion of the Euclidean plane is an affine transformation.
\end{thm}
The following exercise provides more general examples of affine transformations.
\begin{thm}{Exercise}\label{ex:afine-linear}
Show that the following maps of a coordinate plane to itself define affine transformations:
\begin{enumerate}[(a)]
\item\label{ex:afine-linear:shear} \index{shear map}\emph{Shear map} defined by $(x,y)\mapsto (x+k\cdot y,y)$ for a constant~$k$.
\item\label{ex:afine-linear:scaling}
Rescaling defined by $(x,y)\mapsto (a\cdot x,a\cdot y)$ for a constant $a\ne 0$.
\item $x$-scaling and $y$-scaling defined respectively by
\[(x,y)\mapsto (a\cdot x,y),\quad\text{and}\quad(x,y)\mapsto (x,a\cdot y)\]
for a constant $a\ne 0$.
\item\label{affine-general-formula} A transformation defined by
\[(x,y)\mapsto(a\cdot x+b\cdot y+r,c\cdot x+d\cdot y+s)\]
for constants $a,b,c,d,r,s$ such that the matrix $(\begin{smallmatrix}a&b\\c&d\end{smallmatrix})$ is invertible.
\end{enumerate}
\end{thm}
From the fundamental theorem of affine geometry (\ref{thm:fundamental-theorem-of-affine-geometry}), it will follow that any affine transformation can be written in the form \textit{(\ref{affine-general-formula})}.
Recall that points are \index{collinear}\emph{collinear} if they lie on a single line.
\begin{thm}{Exercise}\label{ex:collinear=affine}
Suppose $P\mapsto P'$ is a bijection of the Euclidean plane that maps collinear triples of points to collinear triples.
Show that $P\mapsto P'$ maps noncollinear triples to noncollinear ones.
Conclude that $P\mapsto P'$ is an affine transformation.
\end{thm}
\begin{thm}{Exercise}\label{ex:circle=affine}
Let $\alpha$ be a bijection of the Euclidean plane that maps any circle to a circle.
Show that $\alpha$ is an affine transformation.
\end{thm}
\section{Constructions}
\begin{wrapfigure}{o}{34mm}
\vskip-15mm
\centering
\includegraphics{mppics/pic-287}
\end{wrapfigure}
Let us consider geometric constructions with a ruler and a \index{parallel!tool}\emph{parallel tool};
the latter allows us to draw a line thru a given point parallel to a given line.
(One may think of the tool on the figure.
It consists of two straight edges joined by two arms that can move while remaining parallel to each other.)
By Exercisers~\ref{ex:affine-par}, any construction with these two tools is invariant with respect to affine transformations.
For example,
to solve the following exercise,
it is sufficient to prove that the midpoint of a given segment can be constructed with a ruler and a parallel tool.
\begin{thm}{Exercise}\label{ex:midpoint-affine}
Let $M$ be the midpoint of a segment $[AB]$ in the Euclidean plane.
Assume that an affine transformation sends the points $A$, $B$, and $M$
to $A'$, $B'$, and $M'$ respectively.
Show that $M'$ is the midpoint of~$[A'B']$.
\end{thm}
The following exercise will be used in the proof of Proposition~\ref{prop:affine-linear}.
\begin{thm}{Exercise}\label{ex:R-hom}
Assume that points with coordinates $(0,0)$, $(1,0)$, $(a,0)$, and $(b,0)$ are given.
Using a ruler and a parallel tool, construct points with coordinates $(a+b,0)$, $(a-b,0)$, $(a\cdot b,0)$, and $(\tfrac a b,0)$.
\end{thm}
\begin{thm}{Exercise}\label{ex:center-circ-affine}
Construct the center of a given circle using the ruler and the parallel tool.
\end{thm}
The shear map (described in \ref{ex:afine-linear}\textit{\ref{ex:afine-linear:shear}}) can change angles between lines almost arbitrarily.
This observation can be used to prove the impossibility of some constructions;
here is one example:
\begin{thm}{Exercise}\label{ex:affine-perp}
Show that one cannot construct a line perpendicular to a given line with the ruler and the parallel tool.
\end{thm}
\section{Fundamental theorem of affine geometry}
In this section, we assume knowledge of vector algebra; namely, multiplication by a real number, addition, and the parallelogram rule.
\begin{thm}{Exercise}\label{ex:parallelogram-rule}
Show that affine transformations map parallelograms to parallelograms.
Conclude that if $P\mapsto P'$ is an affine transformation, then
\[\overrightarrow{XY}=\overrightarrow{AB},
\quad\text{if and only if}\quad
\overrightarrow{X'Y'}=\overrightarrow{A'B'}.\]
\end{thm}
\begin{thm}{Proposition}\label{prop:affine-linear}
Let $P\mapsto P'$ be an affine transformation of the Euclidean plane.
Then, for any triple of points $O$, $X$, $P$, we have
\[\overrightarrow{OP}=t\cdot \overrightarrow{OX}
\quad\text{if and only if}\quad
\overrightarrow{O'P'}=t\cdot \overrightarrow{O'X'}.
\eqlbl{eq:OP=tOX}\]
\end{thm}
\parit{Proof.}
Observe that the affine transformations described in Exercise~\ref{ex:afine-linear}, as well as all motions, satisfy the condition \ref{eq:OP=tOX}.
Therefore a given affine transformation $P\mapsto P'$ satisfies \ref{eq:OP=tOX} if and only if its composition with motions and rescalings satisfies \ref{eq:OP=tOX}.
Applying this observation, we can reduce the problem to its partial case.
Namely, we may assume that $O'=O$, $X'=X$, the point $O$ is the origin of a coordinate system, and $X$ has coordinates $(1,0)$.
In this case, $\overrightarrow{OP}=t\cdot \overrightarrow{OX}$ if and only if $P=(t,0)$.
Since $O$ and $X$ are fixed, the transformation maps the $x$-axis to itself.
That is, $P'=(f(t),0)$ for a function $t\mapsto f(t)$,
or, equivalently, $\overrightarrow{O'P'}=f(t)\cdot \overrightarrow{O'X'}$.
It remains to show that
\[f(t)=t
\eqlbl{eq:f(x)=x}\]
for any $t$.
Since $O'=O$ and $X'=X$, we get that $f(0)=0$ and $f(1)=1$.
Furthermore, according to Exercise~\ref{ex:R-hom}, we have that
$f(x\cdot y)=f(x)\cdot f(y)$ and $f(x+y)=f(x)+f(y)$ for any $x,y\in\mathbb{R}$.
By the algebraic lemma (proved below, see \ref{lem:R-auto}), these conditions imply \ref{eq:f(x)=x}.
\qeds
\begin{thm}{Fundamental theorem of affine geometry}\label{thm:fundamental-theorem-of-affine-geometry}\index{Fundamental theorem of affine geometry}
Suppose that an affine transformation $P\mapsto P'$ maps a nondegenerate triangle $OXY$ to a triangle $O'X'Y'$.
Then $\triangle O'X'Y'$ is nondegenerate, and
\[\overrightarrow{OP}=x\cdot\overrightarrow{OX}+y\cdot\overrightarrow{OY}\quad\text{if and only if}\quad\overrightarrow{O'P'}=x\cdot\overrightarrow{O'X'}+y\cdot\overrightarrow{O'Y'}.\]
\end{thm}
\parit{Proof.}
Since an affine transformation maps lines to lines, the triangle
$O'X'Y'$ is nondegenerate.
Consider points $V$ and $W$ defined by
\[\overrightarrow{OV}=x\cdot\overrightarrow{OX},
\quad
\overrightarrow{OW}=y\cdot\overrightarrow{OY}.
\]
{
\begin{wrapfigure}{o}{33mm}
\centering
\vskip-0mm
\includegraphics{mppics/pic-225}
\end{wrapfigure}
Note that
$\overrightarrow{WP}=\overrightarrow{OV}$.
Applying Exercise~\ref{ex:parallelogram-rule} and the proposition, we get
\begin{align*}\overrightarrow{O'P'}&=\overrightarrow{O'W'}+\overrightarrow{W'P'}=
\\
&=\overrightarrow{O'W'}+\overrightarrow{O'V'}=
\\
&=x\cdot\overrightarrow{O'X'}+y\cdot\overrightarrow{O'Y'}.
\end{align*}
\qedsf
}
\begin{thm}{Exercise}\label{ex:affine-continuous}
Show that any affine transformation is continuous.
\end{thm}
The following exercise provides the converse to Exercise \ref{ex:afine-linear}\textit{\ref{affine-general-formula}}.
\begin{thm}{Exercise}\label{ex:affine-coordinates}
Show that any affine transformation can be written in coordinates as
$(x,y)\mapsto(a\cdot x+b\cdot y+r,\ c\cdot x+d\cdot y+s)$
for constants $a,b,c,d,r,s$ such that the matrix $(\begin{smallmatrix}a&b\\c&d\end{smallmatrix})$ is invertible.
\end{thm}
\begin{thm}{Exercise}\label{ex:preserved-circle}
Let $\Gamma$ be a circle with center $O$.
Suppose $\alpha\:P\mapsto P'$ is an affine transformation that maps $\Gamma$ to itself.
Show that $\alpha$ is a motion that fixes $O$;
that is, $O'=O$.
\end{thm}
A bijection from the inversive plane to itself is called an \index{inversive!transformation}\emph{inversive transformation} if it maps circlines to circlines.
By \ref{thm:inverse-cline}, inversions are inversive transformations.
\begin{thm}{Advaneced exercise}\label{ex:inversions-inversive}
Show that any inversive transformation is a composition of inversions and reflections.
\end{thm}
\section{Algebraic lemma}
The following lemma was used in the proof of Proposition~\ref{prop:affine-linear}.
\begin{thm}{Lemma}\label{lem:R-auto}
Assume $f\:\mathbb{R}\to\mathbb{R}$ is a function such that for any $x,y\in\mathbb{R}$ we have
\begin{enumerate}[(a)]
\item\label{lem:R-auto:a} $f(1)=1$,
\item\label{lem:R-auto:b} $f(x+y)=f(x)+f(y)$,
\item\label{lem:R-auto:c} $f(x\cdot y)=f(x)\cdot f(y)$.
\end{enumerate}
Then $f$ is the identity function; that is,
$f(x)=x$ for any $x\in \mathbb{R}$.
\end{thm}
Note that we did not assume that $f$ is continuous.
A function $f$ satisfying these three conditions
is called a \index{field automorphism}\emph{field automorphism}.
Therefore, the lemma states that the identity function is the only automorphism of the field of real numbers.
For the field of complex numbers, the conjugation $z\mapsto\bar z$ (see Section~\ref{sec:complex-conjugation}) gives an example of a nontrivial automorphism.
\begin{thm}{Exercise}\label{ex:f(1)=1}
Suppose that $f\:\mathbb{R}\to\mathbb{R}$ satisfies the condition (\ref{lem:R-auto:c}) in the lemma.
Show that
\begin{enumerate}[(a)]
\item if $f(1)\ne 1$, then $f(x)=0$ for any $x$;
\item if $f(0)\ne 0$, then $f(x)=1$ for any $x$.
\end{enumerate}
\end{thm}
\parit{Proof.}
By \textit{(\ref{lem:R-auto:b})} we have
$f(0)+f(0)=f(0+0)=f(0)$.
Therefore,
\[f(0)=0.
\eqlbl{eq:0=0}\]
Applying \textit{(\ref{lem:R-auto:b})} again, we get that $0=f(0)=f(x)+f(-x)$.
Therefore,
\[f(-x)=-f(x)
\quad
\text{for any}
\quad
x\in \mathbb{R}.
\eqlbl{eq:f-x}\]
Applying \textit{(\ref{lem:R-auto:a})} and \textit{(\ref{lem:R-auto:b})} recursively, we get that
\begin{align*}
f(2)&=f(1)+f(1)=1+1=2;\\
f(3)&=f(2)+f(1)=2+1=3;\\
&\dots
\end{align*}
Together with \ref{eq:f-x},
the latter implies that
$f(n)=n$
for any integer
$n$.
By~\textit{(\ref{lem:R-auto:c})},
$f(m)=f(\tfrac mn)\cdot f(n)$.
Therefore
$$f(\tfrac mn)=\tfrac mn \eqlbl{eq:m/n}$$
for any rational number~$\tfrac mn$.
Assume~$a\ge 0$.
Then the equation $x^2=a$ has a real solution $x\z=\sqrt{a}$.
Suppose $y=f(x)$.
By~\textit{(\ref{lem:R-auto:c})}, $f(a)=y^2$,
and hence $f(a)\ge 0$.
\raggedcolumns\setlength{\multicolsep}{.5mm}
\setlength{\columnseprule}{1pt}
\begin{multicols}{2}
That is,
\[a\ge 0\quad\Longrightarrow\quad f(a)\ge 0.\eqlbl{a>0=>b>0}\]
\columnbreak
By \ref{eq:f-x},
we also get
\[a\le 0\quad \Longrightarrow\quad f(a)\le 0.\eqlbl{a<0=>b<0}\]
\end{multicols}
\setlength{\columnseprule}{0pt}
Now assume $f(a)\ne a$ for some $a\in\mathbb{R}$.
Then there is a rational number $\tfrac{m}{n}$ that lies between $a$ and $f(a)$;
that is,
the numbers
\[x\z=a-\tfrac{m}{n}\quad\text{and}\quad y=f(a)-\tfrac{m}{n}\]
have opposite signs.
By \ref{eq:m/n},
\begin{align*}
y+\tfrac{m}{n}&=f(a)=
\\
&=f(x+\tfrac{m}{n})=
\\
&=f(x)+f(\tfrac{m}{n})=
\\
&=f(x)+\tfrac{m}{n};
\end{align*}
that is, $f(x)=y$.
By \ref{a>0=>b>0} and \ref{a<0=>b<0} the values $x$ and $y$ cannot have opposite signs --- a contradiction.
\qeds
\section{Three theorems}
We discussed several statements in Euclidean geometry that have an affine nature;
that is, their assumptions and conclusions survive under affine transformations.
The examples include
\ref{thm:parallel-point-reflection},
\ref{lem:parallelogram}\ref{lem:parallelogram:midpoint},
\ref{ex:4parallels},
\ref{thm:centroid},
and \ref{ex:midle}.
In this section, we present two more examples of that type.
Suppose that $A$, $B$, and $X$ are distinct points on one line.
Note that
\[\overrightarrow{AX}=t\cdot \overrightarrow{BX},\]
where $t=\pm\frac{AX}{BX}$.
Choose an affine transformation $P\mapsto P'$.
By \ref{prop:affine-linear}, we have
\[\overrightarrow{A'X'}=t\cdot \overrightarrow{B'X'}.\]
It follows that $\frac{A'X'}{B'X'}=\frac{AX}{BX}$;
that is, \textit{if $A$, $B$, and $X$ are distinct points on one line, then the ratio $\frac{AX}{BX}$ is preserved under affine transformations}.
\begin{thm}{Menelaus's theorem}\index{Menelaus's theorem}
Let $ABC$ be a nondegenerate triangle.
Suppose a line $\ell$ crosses the lines $(BC)$, $(CA)$, and $(AB)$ at three distinct points points $A'$, $B'$, and $C'$.
Then
\[\frac{AC'}{BC'}\cdot\frac{BA'}{CA'}\cdot \frac{CB'}{AB'}=1.\]
\end{thm}
As we saw, the ratios $\frac{AC'}{BC'}$, $\frac{BA'}{CA'}$, $\frac{CB'}{AB'}$ are preserved under affine transformations.
Therefore Menelaus's theorem belongs to affine geometry.
However, in the proof we can (and will) use Euclidean-type arguments.
\parit{Proof.}
Let $X$, $Y$ and $Z$ be the footpoints of $A$, $B$ and $C$ on $\ell$.
By the AA similarity condition, $\triangle AXC'\sim \triangle BYC'$.
\begin{figure}[!ht]
\centering
\includegraphics{mppics/pic-227}
\vskip-10mm
\end{figure}
Therefore,
\[\frac{AC'}{BC'}=\frac{AX}{BY}.\]
The same way we get
\[\frac{BA'}{CA'}=\frac{BY}{CZ}
\quad\text{and}\quad
\frac{CB'}{AB'}=\frac{CZ}{AX}.\]
Therefore,
\[\frac{AC'}{BC'}\cdot\frac{BA'}{CA'}\cdot \frac{CB'}{AB'}=\frac{AX}{BY}\cdot \frac{BY}{CZ}\cdot \frac{CZ}{AX}=1.\]
\qedsf
{
\begin{thm}{Exercise}\label{thm:ceva-affine}\label{ex:ceva-affine}
Let $ABC$ be a nondegenerate triangle.
Suppose three distinct points points $A'$, $B'$, and $C'$ lie on the lines $(BC)$, $(CA)$, and $(AB)$ respectively.
Assume that the lines $(AA')$, $(BB')$ and $(CC')$ meet at a point $X$.
\begin{wrapfigure}{r}{31mm}
\centering
\vskip-2mm
\includegraphics{mppics/pic-229}
\end{wrapfigure}
Use Menelaus's theorem to prove the following.
\begin{enumerate}[(a)]
\item \index{Ceva's theorem}\textbf{Ceva's theorem:}
\[\frac{AC'}{BC'}\cdot\frac{BA'}{CA'}\cdot \frac{CB'}{AB'}=1.\]
\item \index{Van Aubel's theorem}\textbf{Van Aubel's theorem:} If $A'$ lies between $B$ and $C$, then
\[\frac{AC'}{BC'}+\frac{AB'}{CB'}=\frac{AX}{A'X}.\]
\end{enumerate}
\end{thm}
}
These three theorems can be formulated in a more exact way by taking the ratios $\frac{AC'}{BC'}$, $\frac{BA'}{CA'}$, and $\frac{CB'}{AB'}$ with sign.
Namely, given distinct points $A$, $B$, and $X$ on one line, let us define the \index{signed ratio}\emph{signed ratio} $(\frac{AX}{BX})\df\pm\frac{AX}{BX}$, where the sign is chosen to meet the equation
\[\overrightarrow{AX}=\left(\frac{AX}{BX}\right)\cdot \overrightarrow{BX}.\]
(Note that $(\frac{AX}{BX})<0$ if and only if $X$ lies between $A$ and $B$.)
The identities
$\left(\tfrac{AC'}{BC'}\right)\cdot\left(\tfrac{BA'}{CA'}\right)\cdot \left(\tfrac{CB'}{AB'}\right)=\pm1$,
with $+1$ and $-1$ give more exact versions of Menelaus's and Ceva's theorem respectively.
In Van Aubel's theorem we have $(\tfrac{AC'}{BC'})+(\tfrac{AB'}{CB'})=(\tfrac{AX}{A'X})$ and this identity is valid without the assumption that $A'$ lies between $B$ and $C$.
We also have the converses of all three signed versions.