Open uniform bound for the competitive finite-game values
OpenKServer.finite_game_uniform_value_boundcompetitive-analysisdynamic-programmingk-serveropen-problem
Fix , an arbitrary metric space , and an initial configuration . In the finite stopping-game recursion, take
The conjecture is that its initial values have one common real upper bound:
The constant may depend on , , and , but on neither the finite request set nor the horizon. The offline optimum is computed in the original ambient metric space from the same initial configuration.
This is an open scalar formulation of KServer.finite_horizon_uniform_bound, using the exact minimax characterization. It is not a proved estimate: the unresolved content is a uniform bound on the iterates of the explicit max-min recursion. A finite value for each individual game, or monotonicity in the horizon, does not establish this assertion.
Preamble
import Definitions.Def_KServer_finite_game open KServer
Formal statement
theorem KServer.finite_game_uniform_value_bound
(k : ℕ) (hk : 1 ≤ k) (M : Type) [MetricSpace M] (C₀ : Config k M) :
∃ a : ℝ, ∀ P : Finset M, ∀ n : ℕ,
finiteGameValue hk P (fun σ => (k : ℝ) * offlineCost C₀ σ) n [] C₀ ≤ a := by sorrySource
Original backward-induction formulation for KServer.finite_horizon_uniform_bound (10ab1333-61a5-4d90-ac6b-b692e2316d1e). The game recursion is specified in this definition, and its exact algorithmic interpretation is proved in KServer.finite_game_characterization. Source of the underlying open conjecture: Koutsoupias, The k-server problem (2009), Conjecture 1, preprint p. 2, https://www.cs.ox.ac.uk/people/elias.koutsoupias/Personal/Papers/paper-kou09.pdf. This recurrence and characterization are an original formal development, not a claim that the conjecture is proved in that reference.