Homotopy types of diffeomorphism groups of polar Morse-Bott foliations on lens spaces 1 Oleksandra Khokhliuk and Sergiy Maksymenko

2025-05-06 0 0 956.34KB 37 页 10玖币
侵权投诉
Homotopy types of diffeomorphism groups of polar
Morse-Bott foliations on lens spaces, 1
Oleksandra Khokhliuk and Sergiy Maksymenko
Abstract. Let T=S1×D2be the solid torus, Fthe Morse-Bott foliation on Tinto 2-tori
parallel to the boundary and one singular circle S1×0, which is the central circle of the
torus T, and D(F, T )the group of diffeomorphisms of Tfixed on T and leaving each leaf
of the foliation Finvariant. We prove that D(F, ∂T )is contractible.
Gluing two copies of Tby some diffeomorphism between their boundaries, we will get
a lens space Lp,q with a Morse-Bott foliation Fp,q obtained from Fon each copy of T. We
also compute the homotopy type of the group D(Fp,q )of diffeomorphisms of Lp,q leaving
invariant each leaf of Fp,q .
1. Introduction
One of the main features and applications of algebraic topology and, especially, homotopy
theory, to real world problems is that very often homotopy invariants of “good” spaces are
discrete and can be described as certain properties “stable under small perturbations”.
On the other hand, the concrete computation of homotopy types can be done in not so
many cases. Actually, the homotopy type of a CW-complex Xis, in principle, determined
by Postnikov tower: one can associate to Xa certain “canonical” representative ¯
Xbeing
homotopy equivalent to X, see [25, §4.3, p. 410]. However, that construction is more theo-
retical, since ¯
Xis usually infinite dimensional, and to construct ¯
Xone need to know all the
homotopy groups of X, which we currently do not completely know even for spheres. Real
computations of homotopy types usually are done by other methods, like theory of coverings
or fibre bundles described in standard textbooks, e.g. [25].
For infinite dimensional spaces, e.g. spaces of maps (infinite dimensional manifolds), the
situation is more complicated. It is known that for a smooth compact manifold Xits group
of Cdiffeomorphisms D(X)endowed with CWhitney topology is a Fr´echet manifold,
and therefore has the homotopy type of a CW-complex, see [64] for discussion of homotopy
types of infinite dimensional manifolds. Homotopy types of path components of D(X)are
completely computed for all compact manifolds with dim X2, [76,9,10,11,20]. There
is also a lot of results about the homotopy types of D(X)for dim X= 3, see e.g. [26,18,28]
and references therein. In higher dimensions it is known very little (just because that every
finitely presented group is the fundamental group of some manifold, so the classification of
compact manifolds contains classification of finitely presented groups), and the results are
2020 Mathematics Subject Classification. 57R30, 57T20.
Key words and phrases. Foliation, diffeomorphism, homotopy type, lens space, solid torus.
1
arXiv:2210.11043v3 [math.AT] 1 Aug 2023
2 OLEKSANDRA KHOKHLIUK AND SERGIY MAKSYMENKO
mostly concern with identifying certain non-trivial elements of homotopy groups, e.g. [62,
75,21,8,43,4], see also good reviews by N. Smolentsev [77] and J. Milnor [61].
Another important class of infinite dimensional spaces are groups of homeomorphisms and
diffeomorphisms preserving leaves of a certain foliation. Their homotopy types are studied
even worse. The most relevant results are obtained in the papers by and K. Fukui and
S. Ushiki [17] and K. Fukui [15]. They studied regular foliations on 3-manifolds with finitely
many Reeb components and proved that the identity path component of the group of foliated
(sending leaves to leaves) diffeomorphisms has the homotopy type of a product of circles, see
Remark 1.1.4 below.
Most of other results related with the structure of diffeomorphism groups of foliations
concern with extensions of results by M. Herman [27], W. Thurston [78], J. Mather [59,60],
D. B. A. Epstein [12] on proving perfectness groups of compactly supported diffeomorphisms
isotopic to the identity. Recall that a group is perfect, if it coincides with its commuta-
tor subgroup G= [G, G]. Then perfectness means triviality of the first homology group
H1(K(G, 1),Z)of the corresponding Eilenberg-MacLane space, and this is where the ho-
motopy theory appears. The technique developed in the latter result by Epstein [12] was
extended by W. Ling [49] to a certain class of groups of homeomorphisms, see also [2].
That allowed in turn to extend the above results on perfectness of diffeomorphism groups,
to groups of leaf-preserving homeomorphisms and diffeomorphisms of nonsingular foliations,
see e.g. [68,70,69,72,73,22,1,47] and references therein.
However for singular foliations their groups of diffeomorphisms are less studied, e.g. [16,
71,55,46]. One of the important class of singular foliations constitute the so-called Morse-
Bott foliations, and in particular, foliations by level sets of Morse-Bott functions. They play
an important role in Hamiltonian dynamics and Poisson geometry, see e.g. [14,79,74,50,
81,57,58,13].
Also in [52] and [55, Theorem 3.7] the second author gave rather wide sufficient conditions
on a vector field Fon a manifold Xunder which the identity path component Did(F)of the
group of orbit-preserving diffeomorphisms of Fis either contractible or homotopy equivalent
to the circle. Given h∈ Did(F)one can associate to each xXthe time αh(x)between x
and h(x)along its orbit of F, see Section 2.6. Such a time αh(x)is not uniquely defined,
however it is shown that certain assumptions on zeros of Fand the fact that h∈ Did(F)
allows to choose αhto be Cand continuous in h.
That result as well as all the technique were applied further to study the homotopy types
of stabilizers and orbits of functions f∈ C(X, R)under the natural “left-right” action of the
product D(X)×D(R)on the space C(X, R)defined by the following rule: (h, φ)·f=φfh.
Notice that this action includes the natural “right” action of the subgroup D(X)×idRon
C(X, R). In [53] there were given wide conditions on funder which the inclusion of “right”
stabilizers and orbits into the corresponding “left-right” stabilizers and orbits are homotopy
equivalences. Moreover, for the case dim X= 2, the homotopy types of “right” stabilizers and
orbits were almost completely computed, see e.g. [54,51,56,45,44], and references therein.
However, the technique developed in the above papers is applicable only to 1-dimensional
foliations. Let us also mention that E. Kudryavtseva [38,39,40,41,42] studied the
homotopy types of spaces of Morse functions on surfaces, and partially rediscovering the
above results obtained a general structure of “right” orbits, see [56] for discussion.
In recent three papers [35,36,37] the present authors developed several techniques
for computations of homotopy types of groups of diffeomorphisms of Morse-Bott and more
DIFFEOMORPHISM GROUPS OF MORSE-BOTT FOLIATIONS ON LENS SPACES 3
general classes of “singular” foliations in higher dimensions. In this paper we present explicit
computations for some simplest Morse-Bott foliations on the solid torus and lens spaces,see
Theorems 1.1.2 and 7.1.2.
1.1. Main result. Let Sr={zC| |z|=r}r[0;1] be the family of concentric circles in
Ctogether with the origin S0= 0,
T={(w, z)C2| |z|= 1,|w| ≤ 1}
=S1×D2
be the solid torus in C2,T2:= S1×S1be its boundary, and F={S1×Sr}r[0,1] be the
partition of Tinto 2-tori parallel to the boundary and the central circle S1×0.
Equivalently, Fis a partition into the level sets of the following Morse-Bott function
f:TR,f(w, z) = |z|2, for which the central circle S1×0is a non-degenerate critical
submanifold.
Say that a diffeomorphism h:TTis F-foliated if for each leaf ω∈ F its image h(ω)
is a (possibly distinct from ω) leaf of Fas well. Also his F-leaf preserving, if h(ω) = ωfor
each leaf ω∈ F. Define the following groups:
• D(T)is the group of diffeomorphisms of T;
• D(F)is the group of F-leaf preserving diffeomorphisms of T;
• Dfix(T, T 2)is the subgroup of D(T)consisting of diffeomorphisms fixed on T2;
• Dfix(F, T 2) := D(F)∩ Dfix(T, T 2).
Endow all those groups with the corresponding CWhitney topologies. Our main result is
the following theorem proved in Section 5:
Theorem 1.1.1.The group Dfix(F, T 2)is weakly contractible, that is weakly homotopy
equivalent to a point (i.e. all its homotopy groups vanish).
Notice further that the homotopy types of the groups D(T)and Dfix(T, T 2)are well
known. In fact, Dfix(T, T 2)is contractible, see [29], while the homotopy type of D(T)can be
described as follows, see Section 6for details. Consider the following subgroup of GL(2,Z):
G:= ε0
m δ|mZ, ε, δ ∈ {±1}.
It naturally acts on R2by linear automorphisms and preserves the integral lattice Z2. Hence
it yields an action on 2-torus T2=R2/Z2by its automorphisms as a Lie group, and so one
can define the semidirect product T:= (R2/Z2)Gcorresponding to that action. Moreover,
every element of Tnaturally extends to a certain F-leaf preserving diffeomorphism of T,
so we get an inclusion e:T ⊂ D(F). It is well known that the inclusion T ⊂ D(T)is a
homotopy equivalence, see Lemma 6.3.1 below. As a consequence of Theorem 1.1.1 we get
the following statement which will be proved in Section 6.4:
Theorem 1.1.2.Each of the following inclusions
Te
→ D(F)⊂ D(T),{idT} ⊂ Dfix(F, T 2)⊂ Dfix(T, T 2)
is a weak homotopy equivalence. In particular, every path component of D(F)is weakly
homotopy equivalent to 2-torus R,π0D(F)
=G, while πkDfix(F, T 2)=0for all k0.
Furthermore, gluing two copies of T(each equipped with the same foliation F) via some
diffeomorphism of T2, we will obtain a lens space Lp,q and a foliation Fp,q on it into two
singular circles and parallel 2-tori. We will call such a foliation polar. Then Theorem 1.1.1
will also allow us to compute the homotopy type of the group D(Fp,q)of Fp,q-leaf preserving
diffeomorphisms. Namely, the following statement holds:
4 OLEKSANDRA KHOKHLIUK AND SERGIY MAKSYMENKO
Theorem 1.1.3.The identity path component of D(Fp,q)is weakly homotopy equivalent
to 2-torus T2, and
π0D(Fp,q) =
G,for L0,1=S1×S2,
Z2Z2,for L1,0=S3, L2,1=RP3,
Z2for p > 2.
Moreover, if p > 2and q2̸=±1(mod p), then the inclusion D(Fp,q)⊂ D(Lp,q)is a weak
homotopy equivalence.
In fact, in Section 5we will obtain a more explicit description of the homotopy type of
D(Fp,q), see Theorem 7.1.2.
Remark 1.1.4.Notice that the above foliations Fand Fp,q are in some sense «dual» to
the ones considered by K. Fukui and S. Ushiki [17] and K. Fukui [15]. Namely, a solid torus
Tadmits a Reeb foliation Rhaving Tas a leaf and being transversal to Fin the interior
of T. Further, let Lp,q be a lens space obtained by gluing two copies T0and T1of Tvia
some diffeomorphism between their boundaries. Then that Reeb foliation Rof each Tigives
a foliation Rp,q of the corresponding lens space Lp,q which has a joint leaf T0=T1with
Fp,q and is transversal to other leaves of Fp,q.
Structure of the paper. For the proof of contractibility of G0=Dfix(F, T 2)in Theo-
rem 1.1.1 we will construct four nested subgroups G4⊂ G3⊂ G2⊂ G1⊂ G0and show that the
inclusions Gi+1 ⊂ Gi,i= 0,1,2,3, are homotopy equivalences, while the smallest group G4is
weakly contractible. Construction of a deformation of Giinto Gi+1,i= 0,1,2,3, requires a
specific technique, and in each case we will establish a more general result which holds for
foliations by level sets of arbitrary smooth and definite homogeneous functions.
Section 2contains several notations and constructions used throughout the paper and also
certain general results about fiberwise definite homogeneous functions on vector bundles. Let
p:EBbe a vector bundle over a smooth manifold Bwhich we identify with its image
in Eunder zero section. In Section 3we prove a “linearization” Theorem 3.1.2 allowing to
isotopy diffeomorphisms of the pair (E, B)to diffeomorphisms coinciding with some vector
bundle morphisms near B. That statement is a particular case of a recent preprint [37] by
the authors, and we present here another and independent proof. It will be used for the proof
of homotopy equivalence G3⊂ G2.
In Section 3.2 we also obtain several consequences of linearization theorem for vector
bundles of rank 1, and, in particular, Lemma 3.2.1 used for the inclusion G1⊂ G0, and
Lemma 3.2.2 used in the proof of Theorem 7.1.2. Also, in Section 3.3, we recall the results
from our paper [36] allowing to prove weak contractibility of G4, see Lemma 3.3.
In Section 4we prove Theorem 4.2.1 for the case B=S1which allows to “unloop diffeo-
morphisms from D(F)along the longitude” and prove homotopy equivalence G2⊂ G1.
The proofs of Theorem 1.1.1 and of the remained inclusion G4⊂ G3are given in Section 5.
In Section 6we recall the description of the homotopy types of groups D(T2),D(T),
Dfix(T, T 2), see Lemmas 6.1.1,6.3.1, and in particular, prove deduce Theorem 1.1.2 from
Theorem 1.1.1, see Section 6.4.
Finally, in Section 7, we compute the homotopy type of D(Fp,q), and prove Theorem 7.1.2
being a detailed variant of Theorem 1.1.3.
DIFFEOMORPHISM GROUPS OF MORSE-BOTT FOLIATIONS ON LENS SPACES 5
2. Preliminaries
For a smooth compact manifold Mwe will denote by Did(M)the identity path component
of the group D(M)of its diffeomorphisms with respect to CWhitney topology. Also the
arrows and will mean monomorphism and epimorphism respectively.
2.1. Deformations. Let Xbe a topological space and H:X×[0; 1] Xbe a homo-
topy. A subset AXwill be called invariant under Hif H(A×[0; 1]) A. Then is Ha
deformation of Xinto Aif H0= idX,H(A×[0; 1]) A, and H1(X)A. It is well known
and is easy to see that in this case the corresponding inclusion map i:AXis a homotopy
equivalence and the map H1:XAis its homotopy inverse.
Moreover, if BXis another subset being invariant under H, then the restriction
H|B×[0;1] :B×[0; 1] Bis a deformation of Binto AB, whence the inclusion ABB
is a homotopy equivalence as well. In that case Hcan be regarded as a deformation of the
pair (X, B)into the pair (A, A B), so the inclusion (A, A B)(X, B)is a homotopy
equivalence of pairs. We will use this observation several times.
2.2. Notation for several diffeomorphism groups. Throughout the paper the term
smooth means Cand all manifolds and diffeomorphisms are assumed to be smooth if the
contrary is not said. Also all spaces of smooth maps, and in particular, diffeomorphism
groups, are endowed with CWhitney topology.
Let Mbe a manifold. Then for each subset BMwe will use the following notations1:
• D(M, B) = {h∈ D(M)|h(B) = B}is the group of diffeomorphisms of Mleaving B
invariant;
• Dfix(M, B)is the group of diffeomorphisms of Mfixed on B;
• Dnb(M, B)is the group of diffeomorphisms of Mfixed on some neighborhood of B
(which may vary for distinct diffeomorphisms).
Moreover, suppose Bis a proper submanifold of M, i.e. B =BM, and p:EBis
a regular neighborhood of B, i.e. a retraction of an open neighborhood EMof Bendowed
with a structure of a vector bundle. Then, in addition,
• D(M, B, p)is the subgroup of D(M, B)consisting of diffeomorphisms hwhich coincide
with some vector bundle morphism ˆ
h:EEon some neighborhood of B(depending on
h); note that Dnb(M, B)⊂ Dfix(M, B, p)since every diffeomorphism hfixed near Bcoincides
thus with the identity automorphism idEof p;
• Dfix(M, B, p) := D(M, B, p)∩ Dfix(M, B);
• D(M, B, p, U)is the subset2of D(M, B, p)consisting of diffeomorphisms hwhich coin-
cide with some vector bundle morphism ˆ
h:EEon a given neighborhood Uof B;
• Dfix(M, B, p, U) := D(M, B, p, U)∩ Dfix(M, B).
Finally, let QMbe a subset and η: [0; 1] Ma path such that η(0), η(1) Q.
Then D0
fix(M, Q, η)will denote the subgroup of Dfix(M, Q)consisting of diffeomorphisms
for which the paths ηand hη(having common ends) are homotopic relatively their ends.
1Those notations are used in different parts of the paper, so the reader may skip this subsection and refer
to it on necessity.
2In general, it is not a group, since Umight not be not invariant under diffeomorphisms from D(M, B, p)
摘要:

HomotopytypesofdiffeomorphismgroupsofpolarMorse-Bottfoliationsonlensspaces,1OleksandraKhokhliukandSergiyMaksymenkoAbstract.LetT=S1×D2bethesolidtorus,FtheMorse-BottfoliationonTinto2-toriparalleltotheboundaryandonesingularcircleS1×0,whichisthecentralcircleofthetorusT,andD(F,∂T)thegroupofdiffeomorphism...

展开>> 收起<<
Homotopy types of diffeomorphism groups of polar Morse-Bott foliations on lens spaces 1 Oleksandra Khokhliuk and Sergiy Maksymenko.pdf

共37页,预览5页

还剩页未读, 继续阅读

声明:本站为文档C2C交易模式,即用户上传的文档直接被用户下载,本站只是中间服务平台,本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私,请立即通知玖贝云文库,我们立即给予删除!

相关推荐

分类:图书资源 价格:10玖币 属性:37 页 大小:956.34KB 格式:PDF 时间:2025-05-06

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务
/ 37
客服
关注