Read-back: FiniteEnergyPlane.two_le_windInf_of_negative_puncture
We work in R4, with coordinates 0,1,2,3 and the maximum norm. The theorem takes the following data and hypotheses.
- H:R4→R has a strictly star-shaped level.
- J is a compatible ξ-structure.
- Γ⊂C is a finite set.
- u:C→R4 is a finite-energy curve on C∖Γ. The values of u on Γ play no role in this hypothesis.
- z0∈Γ.
- Q is a Reeb orbit of H with:
- (negative end at z0) NdEnd(u,ez0,Q,−1), where ez0(s,t)=z0+e−2πse−2πit. Explicitly, there are d,c∈R and b>0 such that, as s→+∞ and uniformly in t∈R,
a(u(ez0(s,t)))+TQs−d→0,ebs(proj(u(ez0(s,t)))−xQ(−TQt+c))→0;
- Q is nondegenerate;
- Q.IsCZAtLeast(3);
- Edλ(C∖Γ,u)=0.
Conclusion. There is an integer k≥2 and an s0∈R such that, for every s≥s0, the loop
t⟼ξC(W(s,−t), ∂sW(s,−t)−λW(s,−t)(∂sW(s,−t))RH(W(s,−t)))
has winding number k. Here W(s,t)=proj(u(ez0(s,t))), so that W(s,−t)=proj(u(z0+e−2πse2πit)). ∂sW(s,t) is the Fréchet derivative of W:R2→R4 at (s,t) applied to (1,0), taken to be 0 where W is not differentiable. The winding-number requirement includes that the loop is continuous, 1-periodic and nowhere zero.
Definitions used
Forms.
- j0(v)=(−v1,v0,−v3,v2).
- ω0(u,v)=u0v1−u1v0+u2v3−u3v2.
- λx(v)=21ω0(x,v).
- XH=j0∇H, with ∇H(x)k=DH(x)ek.
- RH(x)=λx(XH(x))−1XH(x), with 0−1=0.
- ξx={v:λx(v)=0, DH(x)v=0}.
Strictly star-shaped level. H is C∞; every u=0 has a unique r>0 with H(ru)=1; and DH(x)x>0 on {H=1}.
Compatible ξ-structure. J:R4→L(R4) is C∞, and for H(x)=1 and v,w∈ξx:
- J(x)v∈ξx;
- J(x)2v=−v;
- ω0(v,J(x)v)>0 for v=0;
- ω0(J(x)v,J(x)w)=ω0(v,w).
Reeb orbit. Q=(xQ,TQ) with TQ>0, H(xQ)=1, xQ′=RH(xQ), and xQ(τ+TQ)=xQ(τ). The period TQ is not necessarily minimal.
Nondegenerate. For every Φ:R→L(R4) with Φ(0)=id and Φ′(τ)=DRH(xQ(τ))Φ(τ), the only v∈ξxQ(0) with Φ(TQ)v=v is 0.
Radial coordinates. r(y) is a chosen r>0 with H(ry)=1 if one exists, and r(y)=1 otherwise. Then proj(y)=r(y)y and a(y)=−21logr(y).
J~. With x=proj(y), r=r(y), α(v)=2DH(x)xrDH(x)v and β(v)=rλx(v):
J~(y)v=α(v)r−1RH(x)−2β(v)y+J(x)(v−2α(v)y−β(v)r−1RH(x)).
Energies. Write au=a∘u and wu=proj∘u, with ∂sf=Df(z)1 and ∂tf=Df(z)i (both 0 where f is not differentiable).
- Edλ(U,u)=∫Umax(ω0(∂swu,∂twu),0)dz∈[0,∞].
- E(U,u) is the supremum, over C∞ nondecreasing ϕ:R→[0,1], of
∫Umax(ϕ′(au)(∂sauλwu(∂twu)−∂tauλwu(∂swu))+ϕ(au)ω0(∂swu,∂twu),0)dz.
Finite-energy curve on U. This means all of the following:
- u is C∞ on U and nonvanishing there;
- u is differentiable with ∂su+J~(u)∂tu=0 at each z∈U;
- E(U,u)<∞.
Winding number. A loop f:R→C has winding number k if all of the following hold:
- f is continuous, 1-periodic and nowhere zero;
- some continuous θ:R→R has f(t)=∣f(t)∣eiθ(t) for all t;
- θ(1)−θ(0)=2πk.
ξ-coordinates. ξC(x,v)=ω0(v,e2(x))+iω0(e1(x),v), where
e1(x)=∣x∣2−1(q2(x)−DH(x)xDH(x)q2(x)x),e2(x)=∣x∣2−1(q1(x)−DH(x)xDH(x)q1(x)x).
Here q1(x)=(−x2,x3,x0,−x1), q2(x)=(x3,x2,−x1,−x0), and ∣⋅∣2 is the Euclidean norm.
Q.IsCZAtLeast(3). This is a condition on every periodic orbit (y,Ty) of the Hamiltonian system of H that satisfies the two requirements below.
- (orbit) Ty>0, and for all t: y′(t)=(−∂1H,∂0H,−∂3H,∂2H)(y(t)), H(y(t))=1 and y(t+Ty)=y(t).
- (same image and action) {y(τ)}={xQ(τ)} and TQ=∫0Tyλy(τ)(y′(τ))dτ.
For each such orbit and every Y:R→L(R4) with Y(0)=id and Y′(t)=DXH(y(t))Y(t), the condition requires 3≤μCZ(Ψ). Here Ψ and μCZ are defined as follows.
- Ψ(τ)∈R2×2 has (i,j) entry equal to the i-th real component of ξC at y(Tyτ), applied to v−λ(XH)λ(v)XH at y(Tyτ), where v=Y(Tyτ)ej+1(y(0)).
- W is the set of numbers (θ(1)−s)/2π over all s∈[0,2π] and all θ continuous on [0,1] with θ(0)=s such that Ψ(t)(coss,sins) is a positive multiple of (cosθ(t),sinθ(t)) for t∈[0,1].
- With a=infW and b=supW (real infimum and supremum, with junk values if W is empty or unbounded), μCZ(Ψ)=2(⌈b⌉−1) if a≤⌈b⌉−1, and 2⌈b⌉−1 otherwise.
If no such (y,Ty) exists, the condition holds vacuously.
Confirmed by the moderator at approval.