OPERATIONAL FEASIBILITY VALIDATION UNDER UNCERTAINTY FOR DECISION SUPPORT: AN INFERENTIAL PROTOCOL

This paper presents the Two-Stage Inferential Classification Protocol (PCI-2E), an operational and statistical validation layer for patrolling simulations under stochastic uncertainty. For each scenario-policy pair, the protocol declares a critical violation only after three conditions are jointly satisfied: a minimum sample size, a one-sided Wilson lower bound above the operational threshold plus a margin ε, and mandatory retesting with additional seeds for preliminarily critical cases. Every classification is fully traced per pair. In the recalculated baseline, 320 scenario-policy pairs were audited, yielding 305 A (no exceedance, 95.31%), 13 B (inconclusive exceedance, 4.06%), and 2 C-confirmed (critical violation after retesting, 0.63%); six-stage provenance reconciliation identified 14 divergences (4.4%) against the prior snapshot. An external empirical illustration using 2019 field observations from ten Safety Service Patrol beats classified eight beats as A, one as B, and one as C-preliminary; no external case was classified as C-confirmed because independent retest data were unavailable.