1547845439-The_Ricci_Flow_-_Techniques_and_Applications_-_Part_I__Chow_

(jair2018) #1
394 8. APPLICATIONS OF THE REDUCED DISTANCE

for all TE [Ti, T 2 ]). This implies f, is C^00 on M x [Ti, T2] and (8.26) holds on
M x [T 1 , T 2 ]. Thus g ( T) is a shrinking gradient Ricci soliton.
Given To E [T1, T2], by Proposition 1.7, the shrinking gradient Ricci soli-

ton structure ( M, g (To) , \7 f, (To) , - ~ ) may be put into a canonical time-

dependent form (1.11) defined for all t <To,

g(t) = f (t) <p(t)*g (To),
To
where g(t) is a solution of the Ricci fl.ow, by (1.10), i.e., f (t) = To - t
(s = - ~),and <p(t) is a 1-parameter family of diffeomorphisms with <p (0) =
idM. By the uniqueness of complete solutions of the Ricci flow with bounded
curvature (see Chen and Zhu [82]) and since g(O) = g (To), we have
g ( T) = g (To - T) ,
so that
(8.27) <p(To - T)*g (To)= Tog (T) for TE (0, To].
T
Since !Rm [g (T)]I::::; Co< oo for TE [O, T] (we just use this for T small), by
(8.27) we have
sup !Rm [g (To)]!= sup !Rm [<p(To - T)*g (To)]!
M M
T T
::::; - sup IRm [g (T)]I::::; Co-
To M To
for all T E (0,To]. Hence IRm[g(To)]I = 0. Since O(To) = TpM, Mis
diffeomorphic to IR.n. Part (iii) follows since a fl.at shrinking gradient Ricci
soliton on IR.n must be the Gaussian soliton. D
REMARK 8.18. (i) The Riemannian analogue of Corollary 8.17(i) is
lim VolB (p, r) = 1.
r-+O Wnrn
(ii) Note that for the shrinking gradient Ricci soliton g (T) in subsection
7.3 of Chapter 7, the metric g (0) is not well-defined.
The monotonicity of the reduced volume can be easily generalized to the
following. For any fixed measurable subset Ac TpM, we can define D(A, T)

to be the set of vectors V E A such that TV > T, i.e.,

D(A,T) ~{VE A: TV> T} = AnO(T).

It is clear that D(A, T) satisfies D(A, T2) c D(A, T1) if T 2 > T 1.

COROLLARY 8.19 (£-relative volume comparison). Suppose (Mn, g(T)),
T E [O, T], is a complete smooth solution to the backward Ricci flow with the


curvature bound IRm (x, T)I ::::; Co < oo for (x, T) E M x [O, T]. Define for

any T E (0, T) and any measurable subset A c TpM,

VA (T) ~ r (4nT)-nl^2 exp [-f (q, T)] dμg(T) (q).


J Cr exp(D(A,T))

Free download pdf