Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

p. 11, definition of F — F stacks the partial gradients ∇_{x_u} λ_uR_u of the users' costs

Proved
OffloadGNEP.Exist.partial_gradient_eq_F

by mikedeng1 · Oct 8, 2026 · Mathlib 0df444a (Lean v4.33.1)

generalized-nash-equilibriump2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1variational-inequality

Assume Assumption A and the standing hypotheses. Fix a user uuu and a profile xxx with 1−αuxu,m>01-\alpha_ux_{u,m}>01−αu​xu,m​>0 and D(x)=1−1n∑vδvxv,clet>0D(x)=1-\frac1n\sum_v\delta_vx_{v,\mathrm{clet}}>0D(x)=1−n1​∑v​δv​xv,clet​>0. Then the map y↦λuRu(y,x−u)y\mapsto\lambda_uR_u(y,x_{-u})y↦λu​Ru​(y,x−u​) on R3\mathbb R^3R3 is (Fréchet) differentiable at y=xuy=x_uy=xu​, and its gradient is the uuu-th block of FFF:

∇xuλuRu(x)=(αu(1−αuxu,m)2, βu+δu1−1n∑v≠uδvxv,clet(1−1n∑vδvxv,clet)2, γu)⊤=F(x)u.\nabla_{x_u}\lambda_uR_u(x)=\Big(\frac{\alpha_u}{(1-\alpha_ux_{u,m})^2},\ \beta_u+\delta_u\frac{1-\frac1n\sum_{v\ne u}\delta_vx_{v,\mathrm{clet}}}{\big(1-\frac1n\sum_v\delta_vx_{v,\mathrm{clet}}\big)^2},\ \gamma_u\Big)^\top=F(x)_u.∇xu​​λu​Ru​(x)=((1−αu​xu,m​)2αu​​, βu​+δu​(1−n1​∑v​δv​xv,clet​)21−n1​∑v=u​δv​xv,clet​​, γu​)⊤=F(x)u​.

This is what makes FFF the map of the variational inequality associated with the jointly convex game: VI(K,F)\mathrm{VI}(K,F)VI(K,F) collects the first-order conditions of all users at once.

Formalization Note FFF is defined by its closed form, so this statement is not definitional. The derivative is the linear form h↦∑i∈IF(x)u,ihih\mapsto\sum_{i\in I}F(x)_{u,i}h_ih↦∑i∈I​F(x)u,i​hi​ on Tier → ℝ. The two strict inequalities define an open set containing KKK, on which the cost formula has no division by zero.

Preamble
import Mathlib
import Definitions.Def_OffloadGNEP_Exist_Setting
Formal statement
namespace OffloadGNEP.Exist

/-- p. 11, definition of `F`: the `u`-th block of `F` is the partial gradient of user `u`'s
objective `λ_u R_u(·, x_{-u})` with respect to its own variables `x_u`. Stated at every profile `x`
with `1 - α_u x_{u,m} > 0` and `1 - (1/n)∑_v δ_v x_{v,clet} > 0`, an open set containing `K` on
which the closed form of `cost` has no junk division. The derivative is the linear form
`h ↦ ∑_i F(x)_{u,i} h_i` on `ℝ³ = Tier → ℝ`. -/
theorem partial_gradient_eq_F {N : ℕ} (P : Params N) (hA : P.AssumptionA) (hS : P.Standing)
    (u : Fin N) (x : Fin N → Tier → ℝ) (hm : P.alpha u * x u .m < 1) (hD : load P x < 1) :
    HasFDerivAt (fun y : Tier → ℝ => cost P u (Function.update x u y))
      (∑ i : Tier, F P x u i • ContinuousLinearMap.proj (R := ℝ) (φ := fun _ : Tier => ℝ) i)
      (x u) := by sorry

end OffloadGNEP.Exist
Source
Cardellini et al., A game-theoretic approach to computation offloading in mobile cloud computing, accepted manuscript (IRIS Sapienza 11573/779661; DOI 10.1007/s10107-015-0881-6), p. 11, definition of F (display of ∇_{x_u}λ_uR_u and F)
Human review
  • Endorsed by Shuze Chen · Oct 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 8, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me