Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Remark 3.6, p. 174 — a quasi-concave-convex f on [0,1]² with f(·, 1) not u.s.c. has sup inf f = 0 and inf sup f = 1

Proved
SionMinimax.KKM.remark_3_6

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

counterexampleminimaxp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1semicontinuity

Let M=N=[0,1]M = N = [0,1]M=N=[0,1] and define f:M×N→{0,1}f : M \times N \to \{0, 1\}f:M×N→{0,1} by

f(μ,ν)={0if 0≤μ<12 and ν=0, or 12≤μ≤1 and ν=1,1otherwise.f(\mu, \nu) = \begin{cases} 0 & \text{if } 0 \le \mu < \tfrac12 \text{ and } \nu = 0, \text{ or } \tfrac12 \le \mu \le 1 \text{ and } \nu = 1, \\ 1 & \text{otherwise.} \end{cases}f(μ,ν)={01​if 0≤μ<21​ and ν=0, or 21​≤μ≤1 and ν=1,otherwise.​

Then:

  1. for each ν∈[0,1]\nu \in [0,1]ν∈[0,1], f(⋅,ν)f(\cdot, \nu)f(⋅,ν) is quasi-concave on [0,1][0,1][0,1];
  2. for each μ∈[0,1]\mu \in [0,1]μ∈[0,1], f(μ,⋅)f(\mu, \cdot)f(μ,⋅) is quasi-convex on [0,1][0,1][0,1];
  3. for each μ∈[0,1]\mu \in [0,1]μ∈[0,1], f(μ,⋅)f(\mu, \cdot)f(μ,⋅) is lower semicontinuous on [0,1][0,1][0,1];
  4. f(⋅,1)f(\cdot, 1)f(⋅,1) is not upper semicontinuous on [0,1][0,1][0,1];
  5. sup⁡μ∈Minf⁡ν∈Nf(μ,ν)=0\sup_{\mu \in M} \inf_{\nu \in N} f(\mu, \nu) = 0supμ∈M​infν∈N​f(μ,ν)=0 and inf⁡ν∈Nsup⁡μ∈Mf(μ,ν)=1\inf_{\nu \in N} \sup_{\mu \in M} f(\mu, \nu) = 1infν∈N​supμ∈M​f(μ,ν)=1.

The example shows that in Sion's minimax theorem (Theorem 3.4) the upper semicontinuity in the first variable cannot simply be dropped, even for compact convex subsets of the real line: all other hypotheses hold and the minimax equality fails.

Formalization Note The page's condition "0≤μ<1/20\le\mu<1/20≤μ<1/2 and ν=0\nu=0ν=0 or 1/2≤μ≤11/2\le\mu\le11/2≤μ≤1 and ν=1\nu=1ν=1" is read with "and" binding tighter than "or". The suprema and infima are taken over the subtype [0,1][0,1][0,1] and computed in the extended reals EReal through the coercion of the real values. Only the checkable claims about the example are formalized; the informal assertion that the semicontinuity condition "cannot be removed nor appreciably weakened" is not a mathematical statement and is left out.

Preamble
import Mathlib
Formal statement
namespace SionMinimax.KKM
theorem remark_3_6 :
    let f : ℝ → ℝ → ℝ := fun μ ν =>
      if (0 ≤ μ ∧ μ < 1 / 2 ∧ ν = 0) ∨ (1 / 2 ≤ μ ∧ μ ≤ 1 ∧ ν = 1) then 0 else 1
    (∀ ν ∈ Set.Icc (0 : ℝ) 1, QuasiconcaveOn ℝ (Set.Icc (0 : ℝ) 1) (fun μ => f μ ν)) ∧
    (∀ μ ∈ Set.Icc (0 : ℝ) 1, QuasiconvexOn ℝ (Set.Icc (0 : ℝ) 1) (fun ν => f μ ν)) ∧
    (∀ μ ∈ Set.Icc (0 : ℝ) 1, LowerSemicontinuousOn (fun ν => f μ ν) (Set.Icc (0 : ℝ) 1)) ∧
    ¬ UpperSemicontinuousOn (fun μ => f μ 1) (Set.Icc (0 : ℝ) 1) ∧
    (⨆ μ : Set.Icc (0 : ℝ) 1, ⨅ ν : Set.Icc (0 : ℝ) 1, ((f μ ν : ℝ) : EReal)) = 0 ∧
    (⨅ ν : Set.Icc (0 : ℝ) 1, ⨆ μ : Set.Icc (0 : ℝ) 1, ((f μ ν : ℝ) : EReal)) = 1 := by sorry
end SionMinimax.KKM
Source
Sion, On general minimax theorems, Pacific J. Math. 8(1) (1958) 171–176, p. 174 (PDF p. 5), Remark 3.6
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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