Alphabeta Math
TheoremStatement: Literature-sourcedProof: AI-adaptedPipeline-generatedprecheck passjudge pass (gpt-6.1-sol)
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 Hopf--Lax formula solves the Hamilton--Jacobi Cauchy problem

Statement

Let H:Rn→R be finite-valued, continuous, convex, and superlinear, with Legendre transform L, and let u0:Rn→R be bounded and uniformly continuous. Define u(x,t):=Qtu0(x) by The Hopf--Lax operator and the Hopf--Lax formula. Then u is a bounded uniformly continuous function on Rn×[0,T] for every T>0, and: (1) u is a viscosity solution of ut+H(Du)=0 in Rn×(0,∞); (2) u attains the initial datum locally uniformly, sup⁡x∈K∣u(x,t)−u0(x)∣⟶0(t↓0) for every compact K⊆Rn; (3) u is the unique bounded uniformly continuous viscosity solution of the Cauchy problem with datum u0; (4) u satisfies the dynamic-programming relation Qt+su0=QtQsu0 of The Hopf--Lax operators form a semigroup (dynamic programming). No choice principle is used.

Facts & Assumptions

Given: A finite continuous convex superlinear H with Legendre transform L, a bounded uniformly continuous datum u0 with bounded modulus ω (replace any given modulus by its minimum with 2∥u0∥∞), and u=Qtu0.

[F1]

For t>0, Qtu0(x)=inf⁡y{u0(y)+tL((x−y)/t)}, the infimum being attained, and L is real-valued on Rn (The Hopf--Lax operator and the Hopf--Lax formula, Finiteness, superlinearity of the Lagrangian and localisation of Hopf--Lax near-minimisers).

[F2]

Qt+su0=Qt(Qsu0) for all t,s≥0 (The Hopf--Lax operators form a semigroup (dynamic programming)).

[F3]

H=L∗, that is H(p)=sup⁡v(p⋅v−L(v)) (A finite-valued convex Hamiltonian equals its biconjugate).

[F4]

Qtu0 has the modulus ω of u0, and ∣Qtu0(x)−Qtv0(x)∣≤∥u0−v0∥∞ (The Hopf--Lax operator preserves a modulus of continuity, The Hopf--Lax operator is a contraction in the supremum norm).

[F5]

Comparison for autonomous convex superlinear Hamiltonians: bounded uniformly continuous subsolutions and supersolutions of ut+H(Du)=0 on Rn×[0,T] with ordered continuous initial traces satisfy the comparison inequality (Comparison for autonomous convex superlinear Hamiltonians).

Proof

technique · the dynamic-programming inequality in both directions, the biconjugacy $H=L^*$, and comparison for uniqueness
1.1F1F2F4algebra

Boundedness and uniform continuity. For every x and t≥0 we have inf⁡u0−tH(0)≤u(x,t)≤∥u0∥∞+tL(0): the upper bound is the competitor y=x in [F1], and the lower bound follows from L≥−H(0). By [F4] the map x↦u(x,t) is uniformly continuous with modulus ω uniformly in t. For t≥s≥0, [F2] and [F4] give ∣u(x,t)−u(x,s)∣≤∥Qt−su0−u0∥∞; and ∥Qτu0−u0∥∞→0 as τ↓0, since Qτu0≤u0+τL(0) and Qτu0(x)≥u0(x)−sup⁡v{ω(τ∣v∣)−τL(v)}, the supremum tending to 0 as follows. For a>0, set M=∥u0∥∞ and choose A=(2M+1)/a. Superlinearity and continuity of L give b≥0 with L(v)≥A∣v∣−b everywhere. If τ∣v∣≥a, then ω(τ∣v∣)−τL(v)≤−1+bτ; if τ∣v∣<a, then L(v)≥−H(0) gives the bound ω(a)+τH(0). Thus the limsup of the supremum is at most ω(a), which tends to zero as a↓0; its liminf is at least zero by the competitor v=0 and ω(0)=0. Hence u is bounded and uniformly continuous on each strip Rn×[0,T].

1.2F2F3algebra

The subsolution inequality. Let ϕ∈C1 and let u−ϕ have a strict local maximum at (x0,t0) with t0>0. Fix v∈Rn and small h>0, put y:=x0−hv, and use the dynamic-programming identity u(x0,t0)=Qh(Qt0−hu0)(x0)≤u(y,t0−h)+hL(v), the inequality coming from the competitor y in the infimum defining Qh. The contact inequality at (x0,t0) gives u(y,t0−h)≤u(x0,t0)−ϕ(x0,t0)+ϕ(y,t0−h), and combining the two gives ϕ(x0,t0)−ϕ(x0−hv,t0−h)≤hL(v). Dividing by h and letting h↓0 yields ϕt(x0,t0)+⟨Dϕ(x0,t0),v⟩≤L(v) for every v; taking the supremum over v and using H=L∗ of [F3] gives ϕt(x0,t0)+H(Dϕ(x0,t0))≤0. Non-strict maxima are handled by strictification, so u is a viscosity subsolution.

2.1F1F2F3step 1.2algebra

The supersolution inequality. Let u−ϕ have a strict local minimum at (x0,t0) with t0>0. By [F1] there is a minimiser y of u0(y)+t0L((x0−y)/t0); put v:=(x0−y)/t0, so that u(x0,t0)=u0(y)+t0L(v). For 0<h<t0 put zh:=y+t0−ht0(x0−y), so that (zh−y)/(t0−h)=v and zh→x0. The dynamic-programming identity at (zh,t0−h) with the competitor y gives u(zh,t0−h)≤u0(y)+(t0−h)L(v)=u(x0,t0)−hL(v), hence u(x0,t0)−u(zh,t0−h)≥hL(v). The contact inequality at the local minimum gives u(zh,t0−h)≥u(x0,t0)−ϕ(x0,t0)+ϕ(zh,t0−h); combining, ϕ(x0,t0)−ϕ(zh,t0−h)≥hL(v). Dividing by h and letting h↓0 along zh→x0 gives ϕt(x0,t0)+⟨Dϕ(x0,t0),v⟩≥L(v), and since H(Dϕ)=sup⁡w(⟨Dϕ,w⟩−L(w))≥⟨Dϕ,v⟩−L(v) by [F3], we get ϕt+H(Dϕ)≥0. Hence u is a viscosity supersolution, and with step 1.2 it is a viscosity solution of ut+H(Du)=0.

2.2F1step 1.1algebra

The initial trace. For τ>0 and every x, Qτu0(x)≤u0(x)+τL(0) and Qτu0(x)≥u0(x)−sup⁡v{ω(τ∣v∣)−τL(v)}, and the latter supremum tends to 0 by the bounded-modulus estimate in step 1.1; hence sup⁡Rn∣Qτu0−u0∣→0, which is the stated locally uniform (indeed uniform) attainment of the initial datum.

3.1step 1.2step 2.1F2F5∎

Uniqueness and the semigroup. Any bounded uniformly continuous viscosity solution of the Cauchy problem with datum u0 is comparable with u by [F5], in both orders, because both are bounded uniformly continuous and have the same continuous initial trace; hence u is the unique such solution. Property (4) is [F2].

Depends on

Used by

Dependency tree · two levels

30 results within two dependency steps of this one, each drawn at its shortest distance from it. An arrow runs from a result to what uses it, so the chart reads left to right and ends at this result, which carries a heavier outline. Every node is a link to that result. Click elsewhere on the chart to enlarge it.

Sources