Graphics Reference
In-Depth Information
1
2
F £=-+£+
fc
xh
c
xh
+£+
c
e
.
Claim 2.
F and f have the same critical points.
To prove Claim 2, it suffices to prove that the only critical point of F inside U is
0 . But
F
u
F
x
F
h
=
+
x
u
h
u
i
i
i
and
F
(
) <
=- - ¢
1
mx
+
2
h
0
x
F
(
)
=- ¢ +
12
mx
2
h
1.
h
Since —x and —h only vanish at 0 , the same holds for —F and the claim is proved.
F -1 ([c-e,c+e]) is compact and contains no critical points.
Claim 3.
Claim 1 and the fact that F £ f shows that
-
1
(
[
]
) Ã
-
1
(
[
]
)
Fc c
-+
e
,
e
f
c c
-
e
,
+
e
.
It follows that F -1 ([c-e,c-e]) is compact. The only critical point it can contain is 0 , but
this is impossible since
() =- () <-
Fc
m
0
c
e
.
Claim 3 is proved.
Next, define the region H by
-
1
[
]
H
=
F
(
-•-
,
c
e
) -
M
.
c
-
e
Then F -1 ([-•,c-e]) = M c-e » H . Putting all these facts together proves Theorem 8.6.3.
Definition.
The set M c-e » H is usually referred to as M c-e with an attached k-handle
H .
8.6.4. Theorem. Let f : M Æ R be a smooth function that has only nondegenerate
critical points. Assume that M c is compact for all c. Then M has the homotopy type
of a CW complex with one cell of dimension k for each critical point of f with
index k.
Search WWH ::




Custom Search