Alphabeta Math
TheoremStatement: AI-adaptedProof: AI-adaptedPipeline-generatedprecheck pass
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

The endpoint action function of a smooth family of extremals has differential dW=p dq−H dt

Statement

Let n≥1, let J⊆R and U⊆Rn be open and nonempty, and let L:J×U×Rn→R be a fibre-full C2 time-dependent Lagrangian chart that is hyperregular, with inverse Legendre branch v and Hamiltonian H(t,q,p)=p⋅v(t,q,p)−L(t,q,v(t,q,p)) (The Legendre map of a time-dependent Lagrangian: regularity and hyperregularity, A time-dependent Lagrangian chart, its action functional and its fixed-endpoint first variations). Fix t0∈J and q0∈U, let T⊆J∖{t0} be open and nonempty, and let Q⊆U be open and nonempty, let D⊆T×Q×J be open, and let X:D→U be a C2 map such that for every (t,q)∈T×Q the points (t,q,t) and (t,q,t0) lie in D, the closed interval with endpoints t0 and t is contained in the s-section Dt,q:={s∈J:(t,q,s)∈D}, which is required to be an open interval, and:

  1. X(t,q,t)=q and X(t,q,t0)=q0;
  2. s↦X(t,q,s) is a solution of the Euler–Lagrange equation of L on the open interval Dt,q, i.e. (s,X(t,q,s),∂sX(t,q,s))∈J×U×Rn and EL(X(t,q,⋅))=0 on Dt,q.

Define

p(t,q):=Lv(t,q,∂sX(t,q,t)),W(t,q):=∫t0tL(s,X(t,q,s),∂sX(t,q,s))ds((t,q)∈T×Q).

Then W is of class C1 on T×Q and

Wq(t,q)=p(t,q),Wt(t,q)=−H(t,q,p(t,q))((t,q)∈T×Q).

Facts & Assumptions

Given: The chart data L with its inverse branch v and Hamiltonian H, the fixed data t0,q0, the open set D and the C2 family X, and the functions p and W displayed in the Statement.

[F1]

The integrand G(t,q,s):=L(s,X(t,q,s),∂sX(t,q,s)) is of class C1 on D, because L is C2 and X is C2 (Ck Euclidean maps are closed under componentwise algebra and composition, The chain rule for total derivatives: D(g∘f)(a)=Dg(f(a))∘Df(a), Ck Euclidean maps and diffeomorphisms). Mixed second partial derivatives of X commute by Clairaut--Schwarz theorem for continuous second partial derivatives.

[F2]

Integrals with reversed limits mean ∫t0uG=−∫ut0G (The integral with oriented limits: ∫aaf:=0 and ∫baf:=−∫abf). Differentiation of an integral with a variable upper limit and parameters (The first fundamental theorem: if f is integrable on [a,b] and continuous at c, then F′(c)=f(c); in particular a continuous f has F as a primitive, For a<c<b: f is integrable on [a,b] if and only if it is integrable on [a,c] and on [c,b], and then ∫abf=∫acf+∫cbf; with the oriented form for arbitrary a,b,c, Leibniz's rule on a compact rectangle: an interior parameter derivative with a continuous extension may be passed through a Riemann integral, The chain rule for total derivatives: D(g∘f)(a)=Dg(f(a))∘Df(a), If all partial derivatives exist on a neighbourhood and are continuous at a point, then the map is totally differentiable there with Jacobian derivative): for Φ(u,v,q):=∫t0uG(v,q,s) ds one has Φu=G(v,q,u),Φv=∫t0u∂vG(v,q,s) ds,Φq=∫t0u∂qG(v,q,s) ds, locally on a parameter rectangle containing the integration segment. For the upper-limit derivative, additivity writes Φ(u,v,q) as a primitive with a fixed lower endpoint minus its value at t0; for reversed limits the parameter derivatives follow by antisymmetry. The displayed partial derivatives are continuous because their integrands are continuous on a compact rectangle. Since W(t,q)=Φ(t,t,q), the chain rule gives that W is C1 and Wt(t,q)=G(t,q,t)+∫t0t∂tG(t,q,s) ds,Wq(t,q)=∫t0t∂qG(t,q,s) ds.

[F4]

The Euler–Lagrange equation along s↦X(t,q,s) says ∂s(Lv(s,X,∂sX))=Lq(s,X,∂sX) for every (t,q,s)∈D (A time-dependent Lagrangian chart, its action functional and its fixed-endpoint first variations); the notation Lq,Lv is that of The Jacobian matrix of partial derivatives and the gradient in the scalar-valued case and The Euclidean inner product ⟨x,y⟩=∑k<nxkyk on Rn.

[F5]

Endpoint differentiation (The chain rule for total derivatives: D(g∘f)(a)=Dg(f(a))∘Df(a), The derivative and the Riemann integral of a vector-valued function: an intrinsic derivative and a componentwise integral, Intervals of R: the nine order-convex forms, nondegeneracy, and length): the identities X(t,q,t)=q and X(t,q,t0)=q0 hold for every (t,q)∈T×Q, and X is C2, so they may be differentiated in q and in t; this gives ∂qX(t,q,t)=In,∂qX(t,q,t0)=0,∂tX(t,q,t0)=0,∂tX(t,q,t)+∂sX(t,q,t)=0.

Proof

technique · direct; differentiate the endpoint integral in $q$ and in $t$ and use the Euler–Lagrange equation to remove the interior terms
1.1F1F2F3F4algebra

At each (t,q)∈T×Q, the compact integration segment lies in open D. A finite cover by product neighbourhoods yields a common parameter neighbourhood and an open s-interval containing that segment where G is C1; shrinking the parameter neighbourhood also covers nearby upper limits. Thus [F2] applies locally, with the integral oriented if t<t0, and W is C1. Put Jq:=∂qX and P(s):=Lv(s,X,∂sX), both evaluated at fixed (t,q). Using column gradients, Wq=∫t0t[JqTLq+(∂sJq)TP]ds=[JqTP]t0t+∫t0tJqT(Lq−∂sP)ds=[JqTP]t0t. The middle identity is componentwise integration by parts [F3], and the last uses [F4].

1.2F1F2F3F4algebra

For Wt, [F2] gives Wt(t,q)=G(t,q,t)+∫t0t[Lq⋅∂tX+Lv⋅∂s∂tX]ds=L(t,q,∂sX(t,q,t))+[Lv⋅∂tX]s=t0s=t, where the integral was integrated by parts and cancelled by [F4].

2.1F5step 1.1givenalgebra

The boundary terms of step 1.1: differentiating X(t,q,t)=q with respect to q gives ∂qX(t,q,t)=In, so the first term is p(t,q); and differentiating X(t,q,t0)=q0 with respect to q gives ∂qX(t,q,t0)=0, so the second term vanishes. Thus Wq(t,q)=p(t,q).

3.1F5step 1.2givenalgebra∎

The boundary terms of step 1.2: differentiating X(t,q,t0)=q0 with respect to t gives ∂tX(t,q,t0)=0, so the lower term vanishes; and differentiating the identity X(t,q,t)=q with respect to t gives ∂tX(t,q,t)+∂sX(t,q,t)=0. Hence Wt(t,q)=L(t,q,v(t,q))−p(t,q)⋅v(t,q)=−H(t,q,p(t,q)), with v(t,q):=∂sX(t,q,t) and p(t,q)=Lv(t,q,v(t,q)), using H(t,q,p)=p⋅v(t,q,p)−L(t,q,v(t,q,p)) and the inverse relation v(t,q,Lv(t,q,v(t,q)))=v(t,q).

Remarks

  • The two identity clauses determine the first variation of W completely: dW=p dq−H dt at every point of T×Q, where p is the momentum of the selected extremal. This is a conditional statement: it presumes the C2 family X of extremals with the endpoint property, and its proof never asserts that such a family exists.

  • The existence of families with the required endpoint property is the subject of A smooth local family of extremals with prescribed endpoint exists when the endpoint momentum map has invertible derivative, where the hypothesis used is invertibility of the derivative of the endpoint map with respect to the initial momentum. Several extremals may share the same endpoint data while a selected branch still forms a single-valued C2 family. A singular derivative of the endpoint map prevents applying that inverse-function argument, but does not by itself prove that no such family exists. The present theorem applies to every given family satisfying its hypotheses.

  • The identities Wq=p and Wt=−H(t,q,Wq) are the Hamilton–Jacobi relations for the endpoint action on the chart. The second follows from the first by substitution, since p=Wq; the theorem proves the pair directly.

Depends on

Dependency tree · 2 levels

Sources