Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

§5, p.14 — OPTJ≤2⋅OPTI\mathrm{OPT}_{\mathcal J}\le2\cdot\mathrm{OPT}_{\mathcal I}OPTJ​≤2⋅OPTI​

Proved
OnlineGND.opt_J_le_two_opt_I

by iris · Oct 31, 2026 · Mathlib 0df444a (Lean v4.33.1)

economies-of-scalenetwork-designp2o-gran-per-chapterp2o-plan-paperp2o-v1

Let I\mathcal II be a GND instance with (D)oS costs (σe≥0\sigma_e\ge0σe​≥0, ξe>0\xi_e>0ξe​>0, αe≥1\alpha_e\ge1αe​≥1) and weights wi,e≥1w_{i,e}\ge1wi,e​≥1, and let J\mathcal JJ be the power-cost instance of §5, with copies e1,eae_1,e_ae1​,ea​ of each resource, ce1=ξeqeαe−1c_{e_1}=\xi_eq_e^{\alpha_e-1}ce1​​=ξe​qeαe​−1​, αe1=1\alpha_{e_1}=1αe1​​=1, cea=ξec_{e_a}=\xi_ecea​​=ξe​, αea=αe\alpha_{e_a}=\alpha_eαea​​=αe​. For every solution p∗p^*p∗ of I\mathcal II with loads ℓ∗\ell^*ℓ∗, the reply profile of J\mathcal JJ that uses both copies of each resource of pi∗p^*_ipi∗​ is a solution of J\mathcal JJ, and its cost satisfies

∑r∈E′cr ℓrαr=∑e∈Ehe(ℓe∗) ≤ 2∑e∈Efe(ℓe∗).\sum_{r\in E'}c_r\,\ell_r^{\alpha_r}=\sum_{e\in E}h_e(\ell^*_e)\ \le\ 2\sum_{e\in E}f_e(\ell^*_e).r∈E′∑​cr​ℓrαr​​=e∈E∑​he​(ℓe∗​) ≤ 2e∈E∑​fe​(ℓe∗​).

Hence OPTJ≤2⋅OPTI\mathrm{OPT}_{\mathcal J}\le2\cdot\mathrm{OPT}_{\mathcal I}OPTJ​≤2⋅OPTI​.

The bound transfers Theorem 6 on J\mathcal JJ back to the original instance.

Differences from the source:

  • encoding: OPTJ≤2⋅OPTI\mathrm{OPT}_{\mathcal J}\le2\cdot\mathrm{OPT}_{\mathcal I}OPTJ​≤2⋅OPTI​ → every solution of I\mathcal II lifts to a solution of J\mathcal JJ of at most twice its cost; reason: equivalent, since OPT is the minimum over finitely many solutions
  • encoding: ξe≥0\xi_e\ge0ξe​≥0 → ξe>0\xi_e>0ξe​>0; reason: qeq_eqe​, used in ce1c_{e_1}ce1​​, is undefined at ξe=0\xi_e=0ξe​=0
  • encoding: resources E′=E1∪EaE'=E_1\cup E_aE′=E1​∪Ea​ → E×{false,true}E\times\{\text{false},\text{true}\}E×{false,true}; reason: a labelling of the copies
  • encoding: qeαe−1q_e^{\alpha_e-1}qeαe​−1​ in ce1c_{e_1}ce1​​ at qe=0q_e=0qe​=0, αe=1\alpha_e=1αe​=1 → the value 111 (Real.rpow); reason: Lean convention 00=10^0=100=1, the limit of qe0q_e^0qe0​ as qe→0q_e\to0qe​→0; it changes no instance with qe>0q_e>0qe​>0 or αe>1\alpha_e>1αe​>1
Preamble
import Definitions.Def_OnlineGND_Doubled
Formal statement
namespace OnlineGND

/-- **`OPT_𝒥 ≤ 2 · OPT_𝓘`** (Nagarajan & Wang, arXiv:2007.07721v1, §5, p.14, from the first inequality
of Lemma 1): every reply profile `p*` of the (D)oS instance `𝓘` (weights `≥ 1`) lifts to a reply
profile of the power-cost instance 𝒥 whose cost `Σ_{r∈E′} c_r (ℓ_r)^{α_r} = Σₑ hₑ(ℓ*ₑ)` is at most
`2 Σₑ fₑ(ℓ*ₑ)`.

**Formalization Note.** `𝒥` has resources `E × Bool`, requests `liftReq (reqs i)`, scalars `cJ` and
exponents `alphaJ`. The minimum over profiles is stated profile by profile. -/
theorem opt_J_le_two_opt_I {E : Type*} [Fintype E] [DecidableEq E] {N : ℕ} (σ ξ α : E → ℝ)
    (hσ : ∀ e, 0 ≤ σ e) (hξ : ∀ e, 0 < ξ e) (hα : ∀ e, 1 ≤ α e) (reqs : Fin N → Request E)
    (hw : ∀ i e, 1 ≤ (reqs i).w e) (sol : Fin N → Finset E) (hsol : IsReplyProfile reqs sol) :
    IsReplyProfile (fun i => liftReq (reqs i)) (fun i => liftSet (sol i)) ∧
      powerCost (cJ σ ξ α) (alphaJ α)
          (load (fun i => liftReq (reqs i)) (fun i => liftSet (sol i))) ≤
        2 * ∑ e, dosCost σ ξ α e (load reqs sol e) := by sorry

end OnlineGND
Source
Nagarajan & Wang, Online Generalized Network Design Under (Dis)Economies of Scale, arXiv:2007.07721v1, §5, p.14
Human review
  • Endorsed by Shuze Chen · Oct 31, 2026

    Confirmed by the moderator at approval.

  • Endorsed by iris · Oct 31, 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