Each result cited is universally quantified over the data in its own statement. Throughout, ∣a+b∣H2=∣a∣H2+2⟨a,b⟩H+∣b∣H2 for a,b∈H by Elementary Identities in a Real Inner Product Space §expansion, and ∣⟨a,b⟩H∣≤∣a∣H∣b∣H by The Cauchy-Schwarz Inequality in a Real Inner Product Space. We record one identity: for x,y∈D(A),
⟨Ax−Ay,x−y⟩H=∣x−y∣V2,(1)
since x−y∈V, so that ⟨Ax,x−y⟩H=⟨x,x−y⟩V and ⟨Ay,x−y⟩H=⟨y,x−y⟩V by Hilbert Triples: Standing Notation and Background §operator, and subtracting gives ⟨x−y,x−y⟩V.
Claim 1. Each of λ0r, 21∣p∣H2, ⟨Ax,p⟩H and g(x) is a real number for (x,r,p,X)∈D(A)×R×H×Sym(V), so F is a function from D(A)×R×H×Sym(V) to R. Since W=D(A)∩H=D(A), this is exactly the shape required by Second-Order Equation Operator on an Open Subset of a Hilbert Triple and Its δ-Shifts §operator. The defining expression does not involve the fourth argument, so F(x,r,p,X)=F(x,r,p,X′) for all X,X′∈Sym(V), i.e. F is first order; degenerate ellipticity follows from A First-Order Equation Operator is Degenerate Elliptic and Its δ-Shifts Ignore the Form Argument §elliptic.
Claim 2. Let R be positive, let x∈D(A), p∈H, X∈Sym(V) and let r,s∈R satisfy −R≤s≤r≤R. All terms of F other than λ0r are unchanged when r is replaced by s, so
F(x,r,p,X)−F(x,s,p,X)=λ0(r−s),
and in particular λ0(r−s)≤F(x,r,p,X)−F(x,s,p,X). Thus λ0 is a properness constant for F at R, for every positive R, and F is locally strictly proper.
Claim 3. For a real α>1 the function t↦ω2(t,α)=3α2t is the linear function with the nonnegative constant 3α2, hence a modulus of continuity by Linear Moduli of Continuity §modulus. Two preliminary inequalities. First, for x,y∈H, since ∣g(x)−g(y)∣≤∣g(x)∣+∣g(y)∣≤2Cg (claim 5 of Properties of the Absolute Value in an Ordered Field) and ∣g(x)−g(y)∣≤ωg(∣x−y∣H), the last assertion of The Nondecreasing Envelope of a Truncated Modulus of Continuity §majorant gives ∣g(x)−g(y)∣≤ω1(∣x−y∣H). Secondly, for a real α>1 and a nonnegative real s,
αs2+α1−s=α1(αs−21)2+4α3 ≥ 0,
as one checks by expanding α1(αs−21)2=αs2−s+4α1; so s≤αs2+α1, and by The Nondecreasing Envelope of a Truncated Modulus of Continuity §monotone, ω1(s)≤ω1(αs2+α1).
Let R be positive, let x,y∈D(A), r∈R, X,Y∈Sym(H) and let α,δ satisfy 1<α and 0<δ<1; put P=α(x−y). By Second-Order Equation Operator on an Open Subset of a Hilbert Triple and Its δ-Shifts §shifted and the expansion of the norm,
Fδ−(x,r,P,X)=λ0r+λ0δh(x)+21∣P∣H2+δ⟨P,Ax⟩H+2δ2∣Ax∣H2+⟨Ax,P⟩H+δ∣Ax∣H2−g(x),
Fδ+(y,r,P,Y)=λ0r−λ0δh(y)+21∣P∣H2−δ⟨P,Ay⟩H+2δ2∣Ay∣H2+⟨Ay,P⟩H−δ∣Ay∣H2−g(y).
Subtracting and using (1) in the form ⟨Ax,P⟩H−⟨Ay,P⟩H=α∣x−y∣V2,
Fδ−(x,r,P,X)−Fδ+(y,r,P,Y)=λ0δ(h(x)+h(y))+α∣x−y∣V2+Tx+Ty+g(y)−g(x),
where
Tx=δ∣Ax∣H2+2δ2∣Ax∣H2+δ⟨P,Ax⟩H,Ty=δ∣Ay∣H2−2δ2∣Ay∣H2+δ⟨P,Ay⟩H.
Since 2δ2∣Ax∣H2≥0 and δ⟨P,Ax⟩H≥−δ∣P∣H∣Ax∣H, while δ(∣Ax∣H−2∣P∣H)2≥0 gives δ∣Ax∣H2−δ∣P∣H∣Ax∣H≥−4δ∣P∣H2, we get Tx≥−4δ∣P∣H2. Since δ≤1 we have δ−2δ2≥2δ, and 2δ(∣Ay∣H−∣P∣H)2≥0 gives 2δ∣Ay∣H2−δ∣P∣H∣Ay∣H≥−2δ∣P∣H2, so Ty≥−2δ∣P∣H2. Moreover h(x)+h(y)≥0 by The Penalty Function h=21∣⋅∣V2 of a Hilbert Triple: Expansion Identities, Closed Sublevel Sets, Density and Local Bounds §nonneg together with claim 2 of Elementary Arithmetic in an Ordered Field, α∣x−y∣V2≥0, and g(y)−g(x)≥−∣g(x)−g(y)∣ by claim 3 of Properties of the Absolute Value in an Ordered Field. Since ∣P∣H2=α2∣x−y∣H2, we conclude
Fδ−(x,r,P,X)−Fδ+(y,r,P,Y) ≥ −∣g(x)−g(y)∣−43δα2∣x−y∣H2.
By the first preliminary inequality and then the second with s=∣x−y∣H, ∣g(x)−g(y)∣≤ω1(α∣x−y∣H2+α1). For the second term, ∣x−y∣H≤∣x∣H+∣y∣H by The Norm Metric of a Real Inner Product Space: Triangle Inequalities, Limits and Continuity §triangle, (a+b)2≤2a2+2b2 for real a,b because 2a2+2b2−(a+b)2=(a−b)2≥0, and ∣x∣H2≤∣x∣V2=2h(x) by Hilbert Triples: Standing Notation and Background §triple and Monotonicity of Squaring on the Nonnegative Elements of an Ordered Field; hence ∣x−y∣H2≤4(h(x)+h(y))≤4(h(x)+h(y)+1), and multiplying by 43δα2≥0,
43δα2∣x−y∣H2≤3α2δ(h(x)+h(y)+1)=ω2(δ(h(x)+h(y)+1),α).
Therefore
Fδ−(x,r,P,X)−Fδ+(y,r,P,Y) ≥ −ω1(α∣x−y∣H2+α1)−ω2(δ(h(x)+h(y)+1),α).
No restriction on r was used, so (ω1,ω2) is a structure pair for F at every positive R, by The First-Order Structure Condition for a Second-Order Equation Operator on a Hilbert Triple §pair.
Claim 4. Let δ,R satisfy 0<δ<1 and 0<R. Recall from Test Data for a Second-Order Equation Operator on a Hilbert Triple and the Admissible Sets §bounded that an R-bounded datum (x,r,p,X) satisfies h(x)<R, ∣r∣<R and ∣p∣H<R.
(a) An upper bound for Fδ+. For R-bounded η=(y,s,p′,X′), expanding as above,
Fδ+(η)=λ0s−λ0δh(y)+21∣p′∣H2−δ⟨p′,Ay⟩H+2δ2∣Ay∣H2+⟨Ay,p′⟩H−δ∣Ay∣H2−g(y).
Here λ0s≤λ0R, −λ0δh(y)≤0, 21∣p′∣H2≤21R2, −δ⟨p′,Ay⟩H≤R∣Ay∣H, ⟨Ay,p′⟩H≤R∣Ay∣H, 2δ2∣Ay∣H2−δ∣Ay∣H2≤−2δ∣Ay∣H2 and −g(y)≤Cg, so
Fδ+(η)≤λ0R+21R2+Cg+2R∣Ay∣H−2δ∣Ay∣H2.(2)
Since 2δ(∣Ay∣H−δ2R)2≥0 gives 2R∣Ay∣H−2δ∣Ay∣H2≤δ2R2, we obtain Fδ+(η)≤Σ, where Σ=λ0R+21R2+Cg+δ2R2.
(b) A lower bound for Fδ−. For R-bounded ξ=(x,r,p,X), in the same way,
Fδ−(ξ) ≥ −λ0R−R∣Ax∣H−R∣Ax∣H+δ∣Ax∣H2−Cg=δ∣Ax∣H2−2R∣Ax∣H−λ0R−Cg,(3)
using λ0δh(x)≥0, 21∣p∣H2≥0 and 2δ2∣Ax∣H2≥0. Since δ(∣Ax∣H−δR)2≥0 gives δ∣Ax∣H2−2R∣Ax∣H≥−δR2, we obtain Fδ−(ξ)≥Ξ, where Ξ=−δR2−λ0R−Cg.
(c) A bound on ∣Ax∣H for admissible data. Let ξ=(x,r,p,X)∈Sδ,R−. By Test Data for a Second-Order Equation Operator on a Hilbert Triple and the Admissible Sets §admissible there is an R-bounded η with Fδ−(ξ)<R+Fδ+(η)≤R+Σ, so by (3)
δ∣Ax∣H2−2R∣Ax∣H≤K,K=R+Σ+λ0R+Cg (≥0).
As 2R∣Ax∣H≤2δ∣Ax∣H2+δ2R2, this gives 2δ∣Ax∣H2≤K+δ2R2, so ∣Ax∣H≤Λ1, the nonnegative square root of δ2(K+δ2R2).
Let now η=(y,s,p′,X′)∈Sδ,R+. There is an R-bounded ξ with Fδ+(η)>Fδ−(ξ)−R≥Ξ−R, so by (2)
2δ∣Ay∣H2−2R∣Ay∣H≤K′,K′=λ0R+21R2+Cg−Ξ+R (≥0),
and since 2R∣Ay∣H≤4δ∣Ay∣H2+δ4R2 we get ∣Ay∣H≤Λ2, the nonnegative square root of δ4(K′+δ4R2).
(d) The shift modulus. Put Λ=Λ1+Λ2, c=R+2Λ and let ω have value ct+21t2 at a nonnegative t. Then ω is a modulus of continuity: its values are nonnegative, and for positive ϵ every t with 0≤t≤τ, where τ is the smaller of 1 and c+1ϵ, satisfies 21t2≤21t and hence ω(t)≤(c+1)t≤ϵ. Note also that ω is nondecreasing on the nonnegative reals, since c≥0.
Let q∈H and Y∈Sym(H). For ξ=(x,r,p,X)∈Sδ,R− we have Fδ−(x,r,p+q,X+Y)=Fδ−(x,r,p+q,X) by A First-Order Equation Operator is Degenerate Elliptic and Its δ-Shifts Ignore the Form Argument §shifts and claim 1, and, expanding the norm,
Fδ−(x,r,p+q,X)−Fδ−(x,r,p,X)=⟨p+δAx,q⟩H+21∣q∣H2+⟨Ax,q⟩H ≤ (R+Λ+Λ)∣q∣H+21∣q∣H2=ω(∣q∣H),
using ∣p∣H<R, δ<1 and ∣Ax∣H≤Λ1≤Λ. Since ω is nondecreasing, ω(∣q∣H)≤ω(∣q∣H+∥Y∥), which is the first condition of The Shift-Continuity Condition on Admissible Test Data §modulus.
For η=(y,s,p′,X′)∈Sδ,R+ we have Fδ+(y,s,p′+q,X′+Y)=Fδ+(y,s,p′+q,X′), again by A First-Order Equation Operator is Degenerate Elliptic and Its δ-Shifts Ignore the Form Argument §shifts and claim 1, and the same computation gives
Fδ+(y,s,p′+q,X′+Y)−Fδ+(y,s,p′,X′)=⟨p′−δAy,q⟩H+21∣q∣H2+⟨Ay,q⟩H ≥ −(R+Λ+Λ)∣q∣H ≥ −ω(∣q∣H+∥Y∥),
using 21∣q∣H2≥0 and ∣Ay∣H≤Λ2≤Λ. Hence ω is a shift modulus for F at (δ,R), and as δ and R were arbitrary, F satisfies the shift-continuity condition.
Claim 5. By claims 1 to 4 the operator F satisfies all hypotheses on the operator in A Comparison Principle on a Hilbert Triple under the First-Order Structure Condition, and u, v, C satisfy the remaining ones, so A Comparison Principle on a Hilbert Triple under the First-Order Structure Condition §comparison gives u(x)≤v(x) for every x∈V.