Equation (2) — the work function recurrence
ProvedKServerConj.eq_2competitive-analysisk-serveronline-algorithmsp2o-gran-per-chapterp2o-plan-paperp2o-strengthenedp2o-v1work-function-algorithm
Let be a pseudometric space, , a configuration of points and requests. The work function, defined as an offline optimum, satisfies for every configuration of points, and for every and every configuration of points
where and terms with count as . That is:
- for every admissible , and
- equality holds for some admissible ,
where is admissible when or .
The recurrence makes the work function computable online; it is what the row replacement of §2 encodes algebraically (Lemma 2.3).
Differences from the source:
- generalization: a metric space (M, d) → a pseudometric space (
PseudoMetricSpace M); reason: the paper, p.1: "our proof works also when d is a pseudometric"; every metric space is a pseudometric space - generalization: M finite (the analysis assumption, p.1) → M arbitrary; reason: the recurrence holds on every space; the finite case is a special case
- encoding: min over x ∈ X with the convention w_t(Y) = ∞ if |Y| < k → ≤ for every admissible x and = for some admissible x, admissible meaning r_t ∉ X or x = r_t; reason: the excluded terms are exactly those the convention makes ∞
- encoding: k servers and a configuration C₀ → hypotheses 1 ≤ k and C₀.card = k; reason: configurations are k-element sets; for k = 0 no configuration contains a request
- encoding: a finite request sequence r₁, …, r_T → r : ℕ → M, with requests r 1, …, r T; reason: values of r outside 1, …, T enter no quantity of the statement
Preamble
import Mathlib import Definitions.Def_KServerConj_workFn
Formal statement
namespace KServerConj
/-- **Equation (2): the work function recurrence.** Coester, Koutsoupias, Zbysiński, *The k-server
conjecture is true*, arXiv:2609.15979 (version of 2026-11-03, sha256 7ca0380a…), §1.1, printed p.2
(PDF p.3): "the work function can be computed online by dynamic programming, via w₀(X) = D(C₀, X)
and the recurrence w_t(X) = min_{x∈X} {w_{t−1}(X − x + r_t) + d(r_t, x)}, (2) where
X − a + b := (X \ {a}) ∪ {b} and we use the convention w_t(Y) = ∞ if |Y| < k."
**Formalization Note.** The work function `workFn` is the offline optimum (its definition), so
both clauses are theorems. The minimum in (2) is written as "`≤` for every admissible `x`, `=` for
some admissible `x`". A point `x ∈ X` is admissible when `X − x + r_t` has `k` points, i.e.
`r_t ∉ X` or `x = r_t`; the excluded terms are exactly those the paper's convention makes `∞`.
The metric space is arbitrary (possibly infinite) and `d` may be a pseudometric. -/
theorem eq_2 {M : Type*} [PseudoMetricSpace M] [DecidableEq M] (k : ℕ) (hk : 1 ≤ k)
(C₀ : Finset M) (hC₀ : C₀.card = k) (r : ℕ → M) :
(∀ X : Finset M, X.card = k → workFn k C₀ r 0 X = confDist C₀ X) ∧
∀ t : ℕ, 1 ≤ t → ∀ X : Finset M, X.card = k →
(∀ x ∈ X, (r t ∉ X ∨ x = r t) →
workFn k C₀ r t X ≤ workFn k C₀ r (t - 1) (insert (r t) (X.erase x)) + dist (r t) x) ∧
∃ x ∈ X, (r t ∉ X ∨ x = r t) ∧
workFn k C₀ r t X = workFn k C₀ r (t - 1) (insert (r t) (X.erase x)) + dist (r t) x := by sorry
end KServerConj
Source
Coester, Koutsoupias, Zbysiński, The k-server conjecture is true, arXiv:2609.15979 (version of 2026-11-03, sha256 7ca0380a…), §1.1, equation (2), p.2 (pdf p.3)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.