Theorem 1 (Kumar–Seidman): if ρ′ₘ < 1 for every machine, all CAF policies are stable for all initial conditions
ProvedKumarSeidman.CAF.caf_policies_stableConsider a manufacturing system with part types and machines, routes (revisits allowed), input rates , processing times and set-up times . For each buffer let , the prior machine and the modified input rate be defined by (11)–(13): if the prior machine exists and is diconnected with , and otherwise. If the more stringent capacity condition
holds, then all CAF policies are stable for all system initial conditions: for all constants and , every fluid trajectory, from any initial levels and initial set-ups, that obeys the clearing rule (Definition 1) and the clear-a-fraction rule (2) with these constants satisfies
Theorem 1 contrasts with the examples of §III, in which clearing policies are unstable although the usual capacity condition holds: on machines that lie in a common cycle of material flow, (14) budgets for the peak rate at which an upstream machine can release parts, not only for the average rate .
Formalization Note The model, the reading of "nonempty" in Definition 1 as "nonempty or with inflow starting now", and the run structure are those of the imported definition files. The bound on each buffer may depend on the trajectory (and so on the initial state); no bound uniform in the initial state is claimed. Condition (1) is not assumed separately: (14) implies it, since . Set-up times may be zero.
import Mathlib import Definitions.Def_KumarSeidman_CAF_System import Definitions.Def_KumarSeidman_CAF_Trajectory import Definitions.Def_KumarSeidman_CAF_Policy
namespace KumarSeidman.CAF
/-- **Theorem 1** (Kumar–Seidman 1990, p. 294). If the more stringent capacity condition (14),
`ρ'_m < 1` for every machine `m`, holds, then all CAF policies are stable for all system
initial conditions: for all CAF constants `ε_m > 0`, `K_m`, and every trajectory from any
initial state obeying the clearing and clear-a-fraction rules with these constants, every
buffer level is bounded on `[0, ∞)`. -/
theorem caf_policies_stable {P M : ℕ} (S : System P M) (h14 : S.StringentCapacityCondition)
(ε K : Fin M → ℝ) (T : Trajectory S) (hT : T.IsCAF ε K) : T.IsStable := by sorry
end KumarSeidman.CAF
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.