author | Kevin Walker <kevin@canyon23.net> |
Tue, 16 Nov 2010 16:55:49 -0800 | |
changeset 630 | e0093da0d39f |
parent 550 | c9f41c18a96f |
child 793 | adfffac7c138 |
permissions | -rw-r--r-- |
169 | 1 |
%!TEX root = ../../blob1.tex |
98 | 2 |
|
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
3 |
\section{Adapting families of maps to open covers} \label{sec:localising} |
98 | 4 |
|
245
7537032ad5a0
more evmap.tex; also testing using hg from office computer; also
Kevin Walker <kevin@canyon23.net>
parents:
210
diff
changeset
|
5 |
|
272 | 6 |
Let $X$ and $T$ be topological spaces, with $X$ compact. |
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
7 |
Let $\cU = \{U_\alpha\}$ be an open cover of $X$ which affords a partition of |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
8 |
unity $\{r_\alpha\}$. |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
9 |
(That is, $r_\alpha : X \to [0,1]$; $r_\alpha(x) = 0$ if $x\notin U_\alpha$; |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
10 |
for fixed $x$, $r_\alpha(x) \ne 0$ for only finitely many $\alpha$; and $\sum_\alpha r_\alpha = 1$.) |
278
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
11 |
Since $X$ is compact, we will further assume that $r_\alpha = 0$ (globally) |
285
0b3e76167461
small/few changes to famodiff
Kevin Walker <kevin@canyon23.net>
parents:
282
diff
changeset
|
12 |
for all but finitely many $\alpha$. |
245
7537032ad5a0
more evmap.tex; also testing using hg from office computer; also
Kevin Walker <kevin@canyon23.net>
parents:
210
diff
changeset
|
13 |
|
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
14 |
Consider $C_*(\Maps(X\to T))$, the singular chains on the space of continuous maps from $X$ to $T$. |
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
15 |
$C_k(\Maps(X \to T))$ is generated by continuous maps |
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
16 |
\[ |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
17 |
f: P\times X \to T , |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
18 |
\] |
272 | 19 |
where $P$ is some convex linear polyhedron in $\r^k$. |
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
20 |
Recall that $f$ is {\it supported} on $S\sub X$ if $f(p, x)$ does not depend on $p$ when |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
21 |
$x \notin S$, and that $f$ is {\it adapted} to $\cU$ if |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
22 |
$f$ is supported on the union of at most $k$ of the $U_\alpha$'s. |
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
23 |
A chain $c \in C_*(\Maps(X \to T))$ is adapted to $\cU$ if it is a linear combination of |
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
24 |
generators which are adapted. |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
25 |
|
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
26 |
\begin{lemma} \label{basic_adaptation_lemma} |
278
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
27 |
Let $f: P\times X \to T$, as above. |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
28 |
Then there exists |
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
29 |
\[ |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
30 |
F: I \times P\times X \to T |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
31 |
\] |
245
7537032ad5a0
more evmap.tex; also testing using hg from office computer; also
Kevin Walker <kevin@canyon23.net>
parents:
210
diff
changeset
|
32 |
such that |
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
33 |
\begin{enumerate} |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
34 |
\item $F(0, \cdot, \cdot) = f$ . |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
35 |
\item We can decompose $P = \cup_i D_i$ so that |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
36 |
the restrictions $F(1, \cdot, \cdot) : D_i\times X\to T$ are all adapted to $\cU$. |
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
37 |
\item If $f$ has support $S\sub X$, then |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
38 |
$F: (I\times P)\times X\to T$ (a $k{+}1$-parameter family of maps) also has support $S$. |
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
39 |
Furthermore, if $Q$ is a convex linear subpolyhedron of $\bd P$ and $f$ restricted to $Q$ |
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
40 |
has support $S' \subset X$, then |
278
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
41 |
$F: (I\times Q)\times X\to T$ also has support $S'$. |
345 | 42 |
\item Suppose both $X$ and $T$ are smooth manifolds, metric spaces, or PL manifolds, and |
43 |
let $\cX$ denote the subspace of $\Maps(X \to T)$ consisting of immersions or of diffeomorphisms (in the smooth case), |
|
44 |
bi-Lipschitz homeomorphisms (in the metric case), or PL homeomorphisms (in the PL case). |
|
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
45 |
If $f$ is smooth, Lipschitz or PL, as appropriate, and $f(p, \cdot):X\to T$ is in $\cX$ for all $p \in P$ |
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
46 |
then $F(t, p, \cdot)$ is also in $\cX$ for all $t\in I$ and $p\in P$. |
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
47 |
\end{enumerate} |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
48 |
\end{lemma} |
245
7537032ad5a0
more evmap.tex; also testing using hg from office computer; also
Kevin Walker <kevin@canyon23.net>
parents:
210
diff
changeset
|
49 |
|
194 | 50 |
\begin{proof} |
98 | 51 |
Our homotopy will have the form |
52 |
\eqar{ |
|
53 |
F: I \times P \times X &\to& X \\ |
|
54 |
(t, p, x) &\mapsto& f(u(t, p, x), x) |
|
55 |
} |
|
56 |
for some function |
|
57 |
\eq{ |
|
58 |
u : I \times P \times X \to P . |
|
59 |
} |
|
272 | 60 |
|
61 |
First we describe $u$, then we argue that it makes the conclusions of the lemma true. |
|
98 | 62 |
|
272 | 63 |
For each cover index $\alpha$ choose a cell decomposition $K_\alpha$ of $P$ |
64 |
such that the various $K_\alpha$ are in general position with respect to each other. |
|
65 |
If we are in one of the cases of item 4 of the lemma, also choose $K_\alpha$ |
|
66 |
sufficiently fine as described below. |
|
98 | 67 |
|
272 | 68 |
\def\jj{\tilde{L}} |
278
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
69 |
Let $L$ be a common refinement of all the $K_\alpha$'s. |
272 | 70 |
Let $\jj$ denote the handle decomposition of $P$ corresponding to $L$. |
71 |
Each $i$-handle $C$ of $\jj$ has an $i$-dimensional tangential coordinate and, |
|
72 |
more importantly for our purposes, a $k{-}i$-dimensional normal coordinate. |
|
73 |
We will typically use the same notation for $i$-cells of $L$ and the |
|
74 |
corresponding $i$-handles of $\jj$. |
|
98 | 75 |
|
273 | 76 |
For each (top-dimensional) $k$-cell $C$ of each $K_\alpha$, choose a point $p(C) \in C \sub P$. |
278
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
77 |
If $C$ meets a subpolyhedron $Q$ of $\bd P$, we require that $p(C)\in Q$. |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
78 |
(It follows that if $C$ meets both $Q$ and $Q'$, then $p(C)\in Q\cap Q'$. |
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
79 |
Ensuring this is possible corresponds to some mild constraints on the choice of the $K_\alpha$.) |
278
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
80 |
|
272 | 81 |
Let $D$ be a $k$-handle of $\jj$. |
273 | 82 |
For each $\alpha$ let $C(D, \alpha)$ be the $k$-cell of $K_\alpha$ which contains $D$ |
83 |
and let $p(D, \alpha) = p(C(D, \alpha))$. |
|
98 | 84 |
|
85 |
For $p \in D$ we define |
|
86 |
\eq{ |
|
273 | 87 |
u(t, p, x) = (1-t)p + t \sum_\alpha r_\alpha(x) p(D, \alpha) . |
98 | 88 |
} |
272 | 89 |
(Recall that $P$ is a convex linear polyhedron, so the weighted average of points of $P$ |
98 | 90 |
makes sense.) |
91 |
||
273 | 92 |
Thus far we have defined $u(t, p, x)$ when $p$ lies in a $k$-handle of $\jj$. |
93 |
We will now extend $u$ inductively to handles of index less than $k$. |
|
94 |
||
95 |
Let $E$ be a $k{-}1$-handle. |
|
96 |
$E$ is homeomorphic to $B^{k-1}\times [0,1]$, and meets |
|
97 |
the $k$-handles at $B^{k-1}\times\{0\}$ and $B^{k-1}\times\{1\}$. |
|
98 |
Let $\eta : E \to [0,1]$, $\eta(x, s) = s$ be the normal coordinate |
|
99 |
of $E$. |
|
100 |
Let $D_0$ and $D_1$ be the two $k$-handles of $\jj$ adjacent to $E$. |
|
101 |
There is at most one index $\beta$ such that $C(D_0, \beta) \ne C(D_1, \beta)$. |
|
274 | 102 |
(If there is no such index, choose $\beta$ |
273 | 103 |
arbitrarily.) |
104 |
For $p \in E$, define |
|
105 |
\eq{ |
|
106 |
u(t, p, x) = (1-t)p + t \left( \sum_{\alpha \ne \beta} r_\alpha(x) p(D_0, \alpha) |
|
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
107 |
+ r_\beta(x) (\eta(p) p(D_0, \beta) + (1-\eta(p)) p(D_1, \beta)) \right) . |
273 | 108 |
} |
98 | 109 |
|
272 | 110 |
|
274 | 111 |
Now for the general case. |
112 |
Let $E$ be a $k{-}j$-handle. |
|
113 |
Let $D_0,\ldots,D_a$ be the $k$-handles adjacent to $E$. |
|
114 |
There is a subset of cover indices $\cN$, of cardinality $j$, |
|
115 |
such that if $\alpha\notin\cN$ then |
|
116 |
$p(D_u, \alpha) = p(D_v, \alpha)$ for all $0\le u,v \le a$. |
|
117 |
For fixed $\beta\in\cN$ let $\{q_{\beta i}\}$ be the set of values of |
|
118 |
$p(D_u, \beta)$ for $0\le u \le a$. |
|
119 |
Recall the product structure $E = B^{k-j}\times B^j$. |
|
120 |
Inductively, we have defined functions $\eta_{\beta i}:\bd B^j \to [0,1]$ such that |
|
121 |
$\sum_i \eta_{\beta i} = 1$ for all $\beta\in \cN$. |
|
122 |
Choose extensions of $\eta_{\beta i}$ to all of $B^j$. |
|
123 |
Via the projection $E\to B^j$, regard $\eta_{\beta i}$ as a function on $E$. |
|
98 | 124 |
Now define, for $p \in E$, |
194 | 125 |
\begin{equation} |
126 |
\label{eq:u} |
|
98 | 127 |
u(t, p, x) = (1-t)p + t \left( |
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
128 |
\sum_{\alpha \notin \cN} r_\alpha(x) p(D_0, \alpha) |
98 | 129 |
+ \sum_{\beta \in \cN} r_\beta(x) \left( \sum_i \eta_{\beta i}(p) \cdot q_{\beta i} \right) |
130 |
\right) . |
|
194 | 131 |
\end{equation} |
98 | 132 |
|
345 | 133 |
This completes the definition of $u: I \times P \times X \to P$. |
134 |
The formulas above are consistent: for $p$ at the boundary between a $k-j$-handle and |
|
135 |
a $k-(j+1)$-handle the corresponding expressions in Equation \eqref{eq:u} agree, |
|
136 |
since one of the normal coordinates becomes $0$ or $1$. |
|
278
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
137 |
Note that if $Q\sub \bd P$ is a convex linear subpolyhedron, then $u(I\times Q\times X) \sub Q$. |
98 | 138 |
|
139 |
\medskip |
|
140 |
||
274 | 141 |
Next we verify that $u$ affords $F$ the properties claimed in the statement of the lemma. |
98 | 142 |
|
143 |
Since $u(0, p, x) = p$ for all $p\in P$ and $x\in X$, $F(0, p, x) = f(p, x)$ for all $p$ and $x$. |
|
144 |
Therefore $F$ is a homotopy from $f$ to something. |
|
145 |
||
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
146 |
|
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
147 |
\medskip |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
148 |
|
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
149 |
Next we show that for each handle $D$ of $J$, $F(1, \cdot, \cdot) : D\times X \to X$ |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
150 |
is a singular cell adapted to $\cU$. |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
151 |
Let $k-j$ be the index of $D$. |
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
152 |
Referring to Equation \eqref{eq:u}, we see that $F(1, p, x)$ depends on $p$ only if |
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
153 |
$r_\beta(x) \ne 0$ for some $\beta\in\cN$, i.e.\ only if |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
154 |
$x\in \bigcup_{\beta\in\cN} U_\beta$. |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
155 |
Since the cardinality of $\cN$ is $j$ which is less than or equal to $k$, |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
156 |
this shows that $F(1, \cdot, \cdot) : D\times X \to X$ is adapted to $\cU$. |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
157 |
|
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
158 |
\medskip |
274 | 159 |
|
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
160 |
Next we show that $F$ does not increase supports. |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
161 |
If $f(p,x) = f(p',x)$ for all $p,p'\in P$, |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
162 |
then |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
163 |
\[ |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
164 |
F(t, p, x) = f(u(t,p,x),x) = f(u(t',p',x),x) = F(t',p',x) |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
165 |
\] |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
166 |
for all $(t,p)$ and $(t',p')$ in $I\times P$. |
278
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
167 |
Similarly, if $f(q,x) = f(q',x)$ for all $q,q'\in Q\sub \bd P$, |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
168 |
then |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
169 |
\[ |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
170 |
F(t, q, x) = f(u(t,q,x),x) = f(u(t',q',x),x) = F(t',q',x) |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
171 |
\] |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
172 |
for all $(t,q)$ and $(t',q')$ in $I\times Q$. |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
173 |
(Recall that we arranged above that $u(I\times Q\times X) \sub Q$.) |
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
174 |
|
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
175 |
\medskip |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
176 |
|
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
177 |
Now for claim 4 of the lemma. |
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
178 |
Assume that $X$ and $T$ are smooth manifolds and that $f$ is a smooth family of diffeomorphisms. |
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
179 |
We must show that we can choose the $K_\alpha$'s and $u$ so that $F(t, p, \cdot)$ is a |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
180 |
diffeomorphism for all $t$ and $p$. |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
181 |
It suffices to |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
182 |
show that the derivative $\pd{F}{x}(t, p, x)$ is non-singular for all $(t, p, x)$. |
98 | 183 |
We have |
184 |
\eq{ |
|
185 |
% \pd{F}{x}(t, p, x) = \pd{f}{x}(u(t, p, x), x) + \pd{f}{p}(u(t, p, x), x) \pd{u}{x}(t, p, x) . |
|
186 |
\pd{F}{x} = \pd{f}{x} + \pd{f}{p} \pd{u}{x} . |
|
187 |
} |
|
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
188 |
Since $f$ is a family of diffeomorphisms and $X$ and $P$ are compact, |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
189 |
$\pd{f}{x}$ is non-singular and bounded away from zero. |
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
190 |
Also, since $f$ is smooth $\pd{f}{p}$ is bounded. |
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
191 |
Thus if we can insure that $\pd{u}{x}$ is sufficiently small, we are done. |
194 | 192 |
It follows from Equation \eqref{eq:u} above that $\pd{u}{x}$ depends on $\pd{r_\alpha}{x}$ |
98 | 193 |
(which is bounded) |
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
194 |
and the differences amongst the various $p(D_0,\alpha)$'s and $q_{\beta i}$'s. |
98 | 195 |
These differences are small if the cell decompositions $K_\alpha$ are sufficiently fine. |
196 |
This completes the proof that $F$ is a homotopy through diffeomorphisms. |
|
197 |
||
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
198 |
If we replace ``diffeomorphism" with ``immersion" in the above paragraph, the argument goes |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
199 |
through essentially unchanged. |
98 | 200 |
|
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
201 |
Next we consider the case where $f$ is a family of bi-Lipschitz homeomorphisms. |
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
202 |
Recall that we assume that $f$ is Lipschitz in the $P$ direction as well. |
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
203 |
The argument in this case is similar to the one above for diffeomorphisms, with |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
204 |
bounded partial derivatives replaced by Lipschitz constants. |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
205 |
Since $X$ and $P$ are compact, there is a universal bi-Lipschitz constant that works for |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
206 |
$f(p, \cdot)$ for all $p$. |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
207 |
By choosing the cell decompositions $K_\alpha$ sufficiently fine, |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
208 |
we can insure that $u$ has a small Lipschitz constant in the $X$ direction. |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
209 |
This allows us to show that $F(t, p, \cdot)$ has a bi-Lipschitz constant |
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
210 |
close to the universal bi-Lipschitz constant for $f$. |
98 | 211 |
|
275
81d7c550b3da
finshed proof of main lemma in famodiff appendix
Kevin Walker <kevin@canyon23.net>
parents:
274
diff
changeset
|
212 |
Since PL homeomorphisms are bi-Lipschitz, we have established this last remaining case of claim 4 of the lemma as well. |
194 | 213 |
\end{proof} |
98 | 214 |
|
325 | 215 |
\begin{lemma} \label{extension_lemma_c} |
345 | 216 |
Let $\cX_*$ be any of $C_*(\Maps(X \to T))$ or singular chains on the |
217 |
subspace of $\Maps(X\to T)$ consisting of immersions, diffeomorphisms, |
|
218 |
bi-Lipschitz homeomorphisms or PL homeomorphisms. |
|
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
219 |
Let $G_* \subset \cX_*$ denote the chains adapted to an open cover $\cU$ |
276 | 220 |
of $X$. |
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
221 |
Then $G_*$ is a strong deformation retract of $\cX_*$. |
276 | 222 |
\end{lemma} |
223 |
\begin{proof} |
|
285
0b3e76167461
small/few changes to famodiff
Kevin Walker <kevin@canyon23.net>
parents:
282
diff
changeset
|
224 |
It suffices to show that given a generator $f:P\times X\to T$ of $\cX_k$ with |
282
7afacaa87bdb
many small changes to famodiff, mostly prose
Scott Morrison <scott@tqft.net>
parents:
281
diff
changeset
|
225 |
$\bd f \in G_{k-1}$ there exists $h\in \cX_{k+1}$ with $\bd h = f + g$ and $g \in G_k$. |
278
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
226 |
This is exactly what Lemma \ref{basic_adaptation_lemma} |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
227 |
gives us. |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
228 |
More specifically, let $\bd P = \sum Q_i$, with each $Q_i\in G_{k-1}$. |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
229 |
Let $F: I\times P\times X\to T$ be the homotopy constructed in Lemma \ref{basic_adaptation_lemma}. |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
230 |
Then $\bd F$ is equal to $f$ plus $F(1, \cdot, \cdot)$ plus the restrictions of $F$ to $I\times Q_i$. |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
231 |
Part 2 of Lemma \ref{basic_adaptation_lemma} says that $F(1, \cdot, \cdot)\in G_k$, |
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
232 |
while part 3 of Lemma \ref{basic_adaptation_lemma} says that the restrictions to $I\times Q_i$ are in $G_k$. |
276 | 233 |
\end{proof} |
272 | 234 |
|
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
235 |
\medskip |
272 | 236 |
|
278
79f7b1bd7b1a
famodiff.tex maybe finished
Kevin Walker <kevin@canyon23.net>
parents:
277
diff
changeset
|
237 |
%%%%%% Lo, \noop{...} |
277 | 238 |
\noop{ |
239 |
||
272 | 240 |
\medskip |
276 | 241 |
|
242 |
\nn{do we want to keep the following?} |
|
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
243 |
|
277 | 244 |
\nn{ack! not easy to adapt (pun) this old text to continuous maps (instead of homeos, as |
245 |
in the old version); just delete (\\noop) it all for now} |
|
246 |
||
276 | 247 |
The above lemmas remain true if we replace ``adapted" with ``strongly adapted", as defined below. |
248 |
The proof of Lemma \ref{basic_adaptation_lemma} is modified by |
|
249 |
choosing the common refinement $L$ and interpolating maps $\eta$ |
|
250 |
slightly more carefully. |
|
251 |
Since we don't need the stronger result, we omit the details. |
|
252 |
||
253 |
Let $X$, $T$ and $\cU$ be as above. |
|
254 |
A $k$-parameter family of maps $f: P \times X \to T$ is |
|
255 |
{\it strongly adapted to $\cU$} if there is a factorization |
|
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
256 |
\eq{ |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
257 |
P = P_1 \times \cdots \times P_m |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
258 |
} |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
259 |
(for some $m \le k$) |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
260 |
and families of homeomorphisms |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
261 |
\eq{ |
276 | 262 |
f_i : P_i \times X \to T |
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
263 |
} |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
264 |
such that |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
265 |
\begin{itemize} |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
266 |
\item each $f_i$ is supported on some connected $V_i \sub X$; |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
267 |
\item the sets $V_i$ are mutually disjoint; |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
268 |
\item each $V_i$ is the union of at most $k_i$ of the $U_\alpha$'s, |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
269 |
where $k_i = \dim(P_i)$; and |
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
270 |
\item $f(p, \cdot) = g \circ f_1(p_1, \cdot) \circ \cdots \circ f_m(p_m, \cdot)$ |
277 | 271 |
for all $p = (p_1, \ldots, p_m)$, for some fixed $g:X\to T$. |
271
cb40431c8a65
begin to revise families of maps appendix
Kevin Walker <kevin@canyon23.net>
parents:
245
diff
changeset
|
272 |
\end{itemize} |
276 | 273 |
|
277 | 274 |
} |
275 |
% end \noop |
|
276 | 276 |