Distribution (differential geometry) explained

M

is an assignment

x\mapsto\Deltax\subseteqTxM

of vector subspaces satisfying certain properties. In the most common situations, a distribution is asked to be a vector subbundle of the tangent bundle

TM

.

Distributions satisfying a further integrability condition give rise to foliations, i.e. partitions of the manifold into smaller submanifolds. These notions have several applications in many fields of mathematics, including integrable systems, Poisson geometry, non-commutative geometry, sub-Riemannian geometry, differential topology.

Even though they share the same name, distributions presented in this article have nothing to do with distributions in the sense of analysis.

Definition

Let

M

be a smooth manifold; a (smooth) distribution

\Delta

assigns to any point

x\inM

a vector subspace

\Deltax\subsetTxM

in a smooth way. More precisely,

\Delta

consists of a collection

\{\Deltax\subsetTxM\}x

of vector subspaces with the following property: Around any

x\inM

there exist a neighbourhood

Nx\subsetM

and a collection of vector fields

X1,\ldots,Xk

such that, for any point

y\inNx

, span

\{X1(y),\ldots,Xk(y)\}=\Deltay.

The set of smooth vector fields

\{X1,\ldots,Xk\}

is also called a local basis of

\Delta

. These need not be linearly independent at every point, and so aren't formally a vector space basis at every point; thus, the term local generating set can be more appropriate. The notation

\Delta

is used to denote both the assignment

x\mapsto\Deltax

and the subset

\Delta=\amalgx\Deltax\subseteqTM

.

Regular distributions

Given an integer

n\leqm=dim(M)

, a smooth distribution

\Delta

on

M

is called regular of rank

n

if all the subspaces

\Deltax\subsetTxM

have the same dimension

n

. Locally, this amounts to ask that every local basis is given by

n

linearly independent vector fields.

\Delta\subsetTM

of rank

n

(this is actually the most commonly used definition). A rank

n

distribution is sometimes called an

n

-plane distribution, and when

n=m-1

, one talks about hyperplane distributions.

Special classes of distributions

Unless stated otherwise, by "distribution" we mean a smooth regular distribution (in the sense explained above).

Involutive distributions

Given a distribution

\Delta

, its sections consist of vector fields on

M,

forming a vector subspace

\Gamma(\Delta)\subseteq\Gamma(TM)=ak{X}(M)

of the space of all vector fields on

M

. (Notation:

\Gamma(TM)

is the space of sections of

TM.

) A distribution

\Delta

is called involutive if

\Gamma(\Delta)\subseteqak{X}(M)

is also a Lie subalgebra: in other words, for any two vector fields

X,Y\in\Gamma(\Delta)\subseteqak{X}(M)

, the Lie bracket

[X,Y]

belongs to

\Gamma(\Delta)\subseteqak{X}(M)

.

Locally, this condition means that for every point

x\inM

there exists a local basis

\{X1,\ldots,Xn\}

of the distribution in a neighbourhood of

x

such that, for all

1\leqi,j\leqn

, the Lie bracket

[Xi,Xj]

is in the span of

\{X1,\ldots,Xn\}

, i.e.

[Xi,Xj]

is a linear combination of

\{X1,\ldots,Xn\}.

Involutive distributions are a fundamental ingredient in the study of integrable systems. A related idea occurs in Hamiltonian mechanics: two functions

f

and

g

on a symplectic manifold are said to be in mutual involution if their Poisson bracket vanishes.

Integrable distributions and foliations

An integral manifold for a rank

n

distribution

\Delta

is a submanifold

N\subsetM

of dimension

n

such that

TxN=\Deltax

for every

x\inN

. A distribution is called integrable if through any point

x\inM

there is an integral manifold. The base spaces of the bundle

\Delta\subsetTM

are thus disjoint, maximal, connected integral manifolds, also called leaves; that is,

\Delta

defines an n-dimensional foliation of

M

.

Locally, integrability means that for every point

x\inM

there exists a local chart

(U,\{\chi1,\ldots,\chin\})

such that, for every

y\inU

, the space

\Deltay

is spanned by the coordinate vectors
\partial
\partial\chi1

(y),\ldots,

\partial
\partial\chin

(y)

. In other words, every point admits a foliation chart, i.e. the distribution

\Delta

is tangent to the leaves of a foliation. Moreover, this local characterisation coincides with the definition of integrability for a

G

-structures
, when

G

is the group of real invertible upper-triangular block matrices (with

(n x n)

and

(m-n,m-n)

-blocks).

It is easy to see that any integrable distribution is automatically involutive. The converse is less trivial but holds by Frobenius theorem.

Weakly regular distributions

Given any distribution

\Delta\subseteqTM

, the associated Lie flag is a grading, defined as

\Delta(0)\subseteq\Delta(1)\subseteq\ldots\subseteq\Delta(i)\subseteq\Delta(i+1)\subseteq\ldots

where

\Delta(0):=\Gamma(\Delta)

,

\Delta(1):=\langle[\Delta(0),\Delta(0)]

infty(M)}
\rangle
l{C
and

\Delta(i+1):=\langle[\Delta(i),\Delta(0)]

infty(M)}
\rangle
l{C
. In other words,

\Delta(i)\subseteqak{X}(M)

denotes the set of vector fields spanned by the

i

-iterated Lie brackets of elements in

\Gamma(\Delta)

. Some authors use a negative decreasing grading for the definition.

Then

\Delta

is called weakly regular (or just regular by some authors) if there exists a sequence

\{TiM\subseteqTM\}i

of nested vector subbundles such that

\Gamma(TiM)=\Delta(i)

(hence

T0M=\Delta

).[1] Note that, in such case, the associated Lie flag stabilises at a certain point

m\inN

, since the ranks of

TiM

are bounded from above by

rank(TM)=dim(M)

. The string of integers

(rank(\Delta(0)),rank(\Delta(1)),\ldots,rank(\Delta(m)))

is then called the grow vector of

\Delta

.

Any weakly regular distribution has an associated graded vector bundle\mathrm(TM):= T^0 M \oplus \Big(\bigoplus_^ T^M/T^iM \Big) \oplus TM/T^m M.Moreover, the Lie bracket of vector fields descends, for any

i,j=0,\ldots,m

, to a

l{C}infty(M)

-linear bundle morphism

gri(TM) x grj(TM)\togri+j+1(TM)

, called the

(i,j)

-curvature
. In particular, the

(0,0)

-curvature vanishes identically if and only if the distribution is involutive.

Patching together the curvatures, one obtains a morphism

l{L}:gr(TM) x gr(TM)\togr(TM)

, also called the Levi bracket, which makes

gr(TM)

into a bundle of nilpotent Lie algebras; for this reason,

(gr(TM),l{L})

is also called the nilpotentisation of

\Delta

.

The bundle

gr(TM)\toM

, however, is in general not locally trivial, since the Lie algebras

gri(TxM):=

i
T
x
i+1
M/T
x

M

are not isomorphic when varying the point

x\inM

. If this happens, the weakly regular distribution

\Delta

is also called regular (or strongly regular by some authors). Note that the names (strongly, weakly) regular used here are completely unrelated with the notion of regularity discussed above (which is always assumed), i.e. the dimension of the spaces

\Deltax

being constant.

Bracket-generating distributions

A distribution

\Delta\subseteqTM

is called bracket-generating (or non-holonomic, or it is said to satisfy the Hörmander condition) if taking a finite number of Lie brackets of elements in

\Gamma(\Delta)

is enough to generate the entire space of vector fields on

M

. With the notation introduced above, such condition can be written as

\Delta(m)=ak{X}(M)

for certain

m\inN

; then one says also that

\Delta

is bracket-generating in

m+1

steps, or has depth

m+1

.

Clearly, the associated Lie flag of a bracket-generating distribution stabilises at the point

m

. Even though being weakly regular and being bracket-generating are two independent properties (see the examples below), when a distribution satisfies both of them, the integer

m

from the two definitions is the same.

Thanks to the Chow-Rashevskii theorem, given a bracket-generating distribution

\Delta\subseteqTM

on a connected manifold, any two points in

M

can be joined by a path tangent to the distribution.[2] [3]

Examples of regular distributions

Integrable distributions

X

on

M

defines a rank 1 distribution, by setting

\Deltax:=\langleXx\rangle\subseteqTxM

, which is automatically integrable: the image of any integral curve

\gamma:I\toM

is an integral manifold.

k

on

M=Rn

is generated by the first

k

coordinate vector fields
\partial
\partialx1

,\ldots,

\partial
\partialxk
. It is automatically integrable, and the integral manifolds are defined by the equations

\{xi=ci\}i=k+1,\ldots,n

, for any constants

ci\inR

.

\Delta(i)=\Gamma(\Delta)

for every

i

), but it is never bracket-generating.

Non-integrable distributions

M=R3

is given by

\Delta=\ker(\omega)\subseteqTM

, for

\omega=dy-z2dx\in\Omega1(M)

; equivalently, it is generated by the vector fields
\partial
\partialx

+z2

\partial
\partialy
and
\partial
\partialz
. It is bracket-generating since

\Delta(2)=ak{X}(M)

, but it is not weakly regular:

\Delta(1)

has rank 3 everywhere except on the surface

z=0

.

M=R2n+1

is given by

\Delta=\ker(\omega)\subseteqTM

, for

\omega=dz+

n
\sum
i=1

xidyi\in\Omega1(M)

; equivalently, it is generated by the vector fields
\partial
\partialyi
and
\partial
\partialxi

+yi

\partial
\partialz
, for

i=1,\ldots,n

. It is weakly regular, with grow vector

(2n,2n+1)

, and bracket-generating, with

\Delta(1)=ak{X}(M)

. One can also define an abstract contact structures on a manifold

M2n+1

as a hyperplane distribution which is maximally non-integrable, i.e. it is as far from being involutive as possible. An analogue of the Darboux theorem shows that such structure has the unique local model described above.

M=R4

is given by

\Delta=\ker(\omega1)\cap\ker(\omega2)\subseteqTM

, for

\omega1=dz-wdx\in\Omega1(M)

and

\omega2=dy-zdx\in\Omega1(M)

; equivalently, it is generated by the vector fields
\partial
\partialx

+z

\partial
\partialy

+w

\partial
\partialz
and
\partial
\partialw
. It is weakly regular, with grow vector

(2,3,4)

, and bracket-generating. One can also define an abstract Engel structure on a manifold

M4

as a weakly regular rank 2 distribution

\Delta\subseteqTM

such that

\Delta(1)

has rank 3 and

\Delta(2)

has rank 4; Engel proved that such structure has the unique local model described above.[4]

Mk+2

is a rank 2 distribution which is weakly regular and bracket-generating, with grow vector

(2,3,\ldots,k+1,k+2)

. For

k=1

and

k=2

one recovers, respectively, contact distributions on 3-dimensional manifolds and Engel distributions. Goursat structures are locally diffeomorphic to the Cartan distribution of the jet bundles

Jk(R,R)

.

Singular distributions

A singular distribution, generalised distribution, or Stefan-Sussmann distribution, is a smooth distribution which is not regular. This means that the subspaces

\Deltax\subsetTxM

may have different dimensions, and therefore the subset

\Delta\subsetTM

is no longer a smooth subbundle.

In particular, the number of elements in a local basis spanning

\Deltax

will change with

x

, and those vector fields will no longer be linearly independent everywhere. It is not hard to see that the dimension of

\Deltax

is lower semicontinuous, so that at special points the dimension is lower than at nearby points.

Integrability and singular foliations

The definitions of integral manifolds and of integrability given above applies also to the singular case (removing the requirement of the fixed dimension). However, Frobenius theorem does not hold in this context, and involutivity is in general not sufficient for integrability (counterexamples in low dimensions exist).

After several partial results,[5] the integrability problem for singular distributions was fully solved by a theorem independently proved by Stefan[6] [7] and Sussmann.[8] [9] It states that a singular distribution

\Delta

is integrable if and only if the following two properties hold:

\Delta

is generated by a family

F\subseteqak{X}(M)

of vector fields;

\Delta

is invariant with respect to every

X\inF

, i.e.
t
(\phi
X)

*(\Deltay)\subseteq

\Delta
t
\phi
X(y)
, where
t
\phi
X
is the flow of

X

,

t\inR

and

y\indom(X)

.

Similarly to the regular case, an integrable singular distribution defines a singular foliation, which intuitively consists in a partition of

M

into submanifolds (the maximal integral manifolds of

\Delta

) of different dimensions.

The definition of singular foliation can be made precise in several equivalent ways. Actually, in the literature there is a plethora of variations, reformulations and generalisations of the Stefan-Sussman theorem, using different notion of singular foliations according to which applications one has in mind, e.g. Poisson geometry[10] [11] or non-commutative geometry.[12] [13]

Examples

M

, its infinitesimal generators span a singular distribution which is always integrable; the leaves of the associated singular foliation are precisely the orbits of the group action. The distribution/foliation is regular if and only if the action is free.

(M,\pi)

, the image of

\pi\sharp=\iota\pi:T*M\toTM

is a singular distribution which is always integrable; the leaves of the associated singular foliation are precisely the symplectic leaves of

(M,\pi)

. The distribution/foliation is regular If and only if the Poisson manifold is regular.

\rho:A\toTM

of any Lie algebroid

A\toM

defines a singular distribution which is automatically integrable, and the leaves of the associated singular foliation are precisely the leaves of the Lie algebroid. The distribution/foliation is regular if and only if

\rho

has constant rank, i.e. the Lie algebroid is regular. Considering, respectively, the action Lie algebroid

M x ak{g}

and the cotangent Lie algebroid

T*M

, one recovers the two examples above.

References

  1. Tanaka. Noboru. 1970-01-01. On differential systems, graded Lie algebras and pseudo-groups. Kyoto Journal of Mathematics. 10. 1. 10.1215/kjm/1250523814. 2156-2261. free.
  2. Chow. Wei-Liang. 1940-12-01. Über Systeme von liearren partiellen Differentialgleichungen erster Ordnung. Mathematische Annalen. de. 117. 1. 98–105. 10.1007/BF01450011. 121523670 . 1432-1807.
  3. Rashevsky. P. K.. 1938. Any two points of a totally nonholonomic space may be connected by an admissible line.. Uch. Zap. Ped. Inst. Im. Liebknechta, Ser. Phys. Math.. Russian. 2. 83–94.
  4. Engel. Friedrich. 1889. Zur Invariantentheorie der Systeme Pfaff'scher Gleichungen. Leipz. Ber.. German. 41. 157–176.
  5. Lavau. Sylvain. 2018-12-01. A short guide through integration theorems of generalized distributions. Differential Geometry and Its Applications. en. 61. 42–58. 10.1016/j.difgeo.2018.07.005. 1710.01627 . 119669163 . 0926-2245.
  6. Stefan. P.. 1974. Accessibility and foliations with singularities. Bulletin of the American Mathematical Society. en. 80. 6. 1142–1145. 10.1090/S0002-9904-1974-13648-7. 0002-9904. free.
  7. Stefan. P.. 1974. Accessible Sets, Orbits, and Foliations with Singularities. Proceedings of the London Mathematical Society. en. s3-29. 4. 699–713. 10.1112/plms/s3-29.4.699. 1460-244X.
  8. Sussmann. Hector J.. 1973. Orbits of families of vector fields and integrability of systems with singularities. Bulletin of the American Mathematical Society. en. 79. 1. 197–199. 10.1090/S0002-9904-1973-13152-0. 0002-9904. free.
  9. Sussmann. Héctor J.. 1973. Orbits of families of vector fields and integrability of distributions. Transactions of the American Mathematical Society. en. 180. 171–188. 10.1090/S0002-9947-1973-0321133-2. 0002-9947. free.
  10. Androulidakis. Iakovos. Zambon. Marco. 2016-04-28. Stefan–Sussmann singular foliations, singular subalgebroids and their associated sheaves. International Journal of Geometric Methods in Modern Physics. 13. Supp. 1. 1641001–1641267. 10.1142/S0219887816410012. 2016IJGMM..1341001A . 0219-8878.
  11. Laurent-Gengoux. Camille. Lavau. Sylvain. Strobl. Thomas. 2020. The Universal Lie ∞-Algebroid of a Singular Foliation. ELibM – Doc. Math.. en. 25. 2020 . 1571–1652. 10.25537/dm.2020v25.1571-1652.
  12. Debord. Claire. 2001-07-01. Holonomy Groupoids of Singular Foliations. Journal of Differential Geometry. 58. 3. 10.4310/jdg/1090348356. 54714044 . 0022-040X. free.
  13. Androulidakis. Iakovos. Skandalis. Georges. 2009-01-01. The holonomy groupoid of a singular foliation. Journal für die reine und angewandte Mathematik (Crelle's Journal) . en. 2009. 626. 1–37. 10.1515/CRELLE.2009.001. 14450917 . 1435-5345. math/0612370.

Books, lecture notes and external links