Lie Groups and Lie Algebras Lecture 5 notes

Left and Right Invariant Vector Fields

Definition: Preliminaries

Consider a smooth manifold MM with a smooth vector field ξ\xi. Then, the vector field defines a system of ODEs on M: ddtx=ξ(t)\frac{d}{dt}x = \xi(t)
A smooth curve γ(t),t(ϵ,ϵ)\gamma(t), t \in (-\epsilon, \epsilon) is an integral curve of ξ\xi if ddtγ(t)=ξ(γ(t))\frac{d}{dt}\gamma(t) = \xi(\gamma(t)) (in other words, γ(t)\gamma(t) is a solution of ddtx=ξ(t)\frac{d}{dt}x = \xi(t))
Existence and uniqueness theorem: For any xMx \in M there is a unique integral curve γx(t)\gamma_x(t) passing through it (s.t. γx(0)=x\gamma_x(0) = x).
Flow ϕt\phi_t: To each vector field ξ\xi, we can assign (at least locally) a diffeomorphism ϕt\phi_t which shifts each point xx along ξ\xi by time tt. In other words, ϕt(x)=γx(t)\phi_t(x) = \gamma_x(t).
"Locally" means that usually the flow ϕt\phi_t is defined only for sufficiently small tt (depends on xx).
Completeness: If each integral curve γx(t)\gamma_x(t) can be extended (in the sense of tt) to R=(,)\mathbb{R} = (-\infty, \infty), then ξ\xi is called complete. Equivalently, completeness of ξ\xi means that the flow ϕt:MM\phi_t: M \to M is globally defined for all tRt \in \mathbb{R}.
Since ϕtphis=ϕt+s\phi_t\circ phi_s = \phi_{t + s}, the flow can be thought of as a group under composition.
The differential of a smooth map F:MNF: M \to N is the map dF:TMTNdF: TM \to TN defined by dF(ddtγ(t))=ddtF(γ(t))dF(\frac{d}{dt}\gamma(t)) = \frac{d}{dt}F(\gamma(t)). (sends tangent vectors to tangent vectors)
Lie bracket of vector fields: Given smooth vector fields ξ\xi and η\eta on MM, we can introduce a new vector field [ξ,η][\xi, \eta] s.t.:
- [ξ,η]k=ξiddxiηkηiddxiξk[\xi, \eta]^k = \xi^i \frac{d}{d x^i} \eta^k - \eta^i \frac{d}{d x^i} \xi^k (in local coords)
- [ξ,η](f)=ξ(η(f))η(ξ(f))[\xi, \eta](f) = \xi(\eta(f)) - \eta(\xi(f))
.#-> This lie bracket represents taking the differential of ξ\xi with respect to η\eta, or is it the other way around?
If F:MNF: M \to N is a diffeomorphism, with dF:TMTNdF: TM \to TN its differential, then dF([ξ,η])=[dF(ξ),dF(η)]dF([\xi, \eta]) = [dF(\xi), dF(\eta)] #-> This is obvious because FF is a transformation of local coords and the Lie bracket is independent on local coords?
Vector fields ξ,η\xi, \eta commute (i.e. [ξ,η]=0[\xi, \eta] = 0) if and only if the corresponding flows commute (i.e. ϕηϕξ=ϕξϕη\phi_{\eta}\circ \phi_{\xi} = \phi_{\xi}\circ \phi_{\eta})

Throughout, we'll use the notation La:GG,xaxL_a: G \to G, x \mapsto ax (left translation by aGa \in G) and Ra:GG,xxaR_a: G \to G, x \mapsto xa (right translation by aGa \in G)
Clearly LaL_a and RaR_a are diffeomorphisms of GG onto itself
We can use dLa:TGTGdL_a: TG \to TG and dRa:TGTGdR_a: TG \to TG for the differentials, and use the same notation for the differential at some fixed point xGx \in G as well.
Left and right translations commute: LaRb=xaxb=RbLaL_a\circ R_b = x \mapsto axb = R_b\circ L_a
But in general, LaLbLbLaL_a\circ L_b \neq L_b\circ L_a and RaRbRbRaR_a\circ R_b \neq R_b\circ R_a
also LbLa=LbaL_b\circ L_a = L_{ba} and RbRa=RbaR_b\circ R_a = R_{ba}
The same applies to differentials

Definition: Left Invariance

A vector field ξ\xi is called left invariant if it is preserved under left translations.
In other words, for any aGa \in G, dLa(ξ(x))=ξ(La(x))dL_a(\xi(x)) = \xi(L_a(x)).
Similarly, a vector field η\eta is called right invariant if dRa(η)=ηdR_a(\eta) = \eta for any aGa \in G.
In other words, if we consider the values ξ(x)\xi(x) and ξ(y)\xi(y) of our vector field ξ\xi at two distinct points xGx \in G and y=axGy = ax \in G, then dLa(ξ(x))=ξ(ax)dL_a(\xi(x)) = \xi(ax). Similar for right invariance.

Theorem: Left-invariant Construction

Take an arbitrary vector ξ0=ξ(e)TeG\xi_0 = \xi(e) \in T_e G at the identity eGe \in G, and define a tangent vector ξ(a)TaG\xi(a) \in T_a G at any other point aGa \in G by putting ξ(a)=dLa(ξ0)\xi(a) = dL_a(\xi_0).
Then, we get a tangent vector ξ(a)\xi(a) for any aGa \in G. ξ\xi is smooth because LaL_a depends on aGa \in G smoothly.
Such a vector field is left-invariant

Proof:
We only need to verify the condition that dLa(ξ(x))=ξ(ax)dL_a(\xi(x)) = \xi(ax) for any x,aGx, a \in G. Notice that for x=ex = e, this condition holds by construction.
For any other xGx \in G, we have dLa(ξ(x))=dLa(dLx(ξ0))=dLadLx(ξ0)=dLax(ξ0)=ξ(ax)dL_a(\xi(x)) = dL_a(dL_x(\xi_0)) = dL_a\circ dL_x(\xi_0) = dL_{ax}(\xi_0) = \xi(ax)

Corollary:
A left invariant vector field ξ\xi is uniquely defined by its "initial" value ξ0=ξ(e)\xi_0 = \xi(e) at the identity. Moreover, ξ0\xi_0 can be chosen arbitrarily

Corollary:
The set of left invariant vector fields is a vector space of dimension dim G\text{dim } G, which is naturally isomorphic to the tangent space TeGT_e G to GG at the identity ee.
This isomorphism is established by the construction ξ(a)=dLa(ξ0)\xi(a) = dL_a(\xi_0)

Properties:
If ξ\xi is a left invariant vector field on GG, then
Proposition 1:
Let γe(t)\gamma_e(t) be the integral curve of ξ\xi passing through the identity ee (i.e. γe(0)=e\gamma_e(0) = e). Then, the integral curve γx(t)\gamma_x(t) of ξ\xi passing through xx is xγe(t)=Lx(γe(t))x \gamma_e(t) = L_x(\gamma_e(t)).
Proof:
ddtLx(γe(t))=dLx(ddtγe(t))\frac{d}{dt} L_x(\gamma_e(t)) = dL_x(\frac{d}{dt} \gamma_e(t)) by def of differential, dLx(ddtγe(t))=dLx(ξ(γe(t))dL_x(\frac{d}{dt} \gamma_e(t)) = dL_x(\xi(\gamma_e(t)) by γe(t)\gamma_e(t) being an integral curve of ξ\xi, and = ξ(Lx(γe(t))\xi(L_x(\gamma_e(t)) by ξ\xi being left-invariant.
Thus, Lx(γe(t))=xγe(t)L_x(\gamma_e(t)) = x \gamma_e(t) is an integral curve of ξ\xi. It suffices that xγe(0)=xe=xx \gamma_e(0) = xe = x to show that xγe(t)x \gamma_e(t) passes through xx.
Corollary:
The left translation of any integral curve of xixi is again an integral curve
Corollary:
The flow ϕξt:GG\phi_{\xi}^t: G \to G of xixi is defined by ϕξt(x)=xγe(t)\phi_{\xi}^t(x) = x \gamma_e(t), where γe(t)\gamma_e(t) is the integral curve of ξ\xi through the identity.
Proposition 2:
ξ\xi is complete, i.e. the flow ϕξt:GG\phi_{\xi}^t: G \to G of ξ\xi is well defined for all tRt \in \mathbb{R}
Proof:
The definition of the flow for a left-invariant vector field ξ\xi is defined as ϕξt(x)=xγe(t)\phi_{\xi}^t(x) = x \gamma_e(t). If γe(t)\gamma_e(t) is defined on (ϵ,ϵ)(-\epsilon, \epsilon), then ϕξt\phi_{\xi}^t is defined on the whole group for t(ϵ,ϵ)t \in (-\epsilon, \epsilon).
Then, ϕt\phi^t can naturally be defined for all t(,)t \in (-\infty, \infty) just by iterating: ϕξt=ϕξtkϕξtk\phi_{\xi}^t = \phi_{\xi}^{\frac{t}{k}}\circ\dots\circ\phi_{\xi}^{\frac{t}{k}} kk times where tk(ϵ,ϵ)\frac{t}{k} \in (-\epsilon, \epsilon).

Definition: One-parameter Subgroup

Definition: A smooth map f:RGf: \mathbb{R} \to G is called a one-parameter subgroup of GG if f(s+t)=f(s)f(t)f(s + t) = f(s)f(t) for any t,sRt, s \in \mathbb{R} #-> i.e. it can be parameterized by R\mathbb{R}