Every zero-index Gittins policy attains the full-information reward
ProvedPandoraBO.Private20261023.zero_gittins_full_informationoptimal-searchpandoras-box
Every policy obeying the zero-multiplier Gittins rule has expected reward equal to the full-information reward and is optimal at zero multiplier. This does not assert that every zero-index tie-breaking rule satisfies a particular budget.
Preamble
import Definitions.Def_PandoraBO_Model set_option autoImplicit false
Formal statement
theorem PandoraBO.Private20261023.zero_gittins_full_information {n : ℕ} (M : PandoraBO.Instance n) (π : PandoraBO.Policy n)
(hπ : PandoraBO.isZeroGittinsPolicy M π) :
PandoraBO.expectedReward M π = PandoraBO.fullInformationReward M ∧
PandoraBO.optimal M 0 π := by sorry
Source
Auxiliary result supporting the zero-multiplier branch in Xie et al., author revised manuscript reviewed 23 October 2026, Appendix B.5.