Proposition 8.5.1 — average cost of serving at a constant rate a > p
ProvedSennottDP.AvgASM.service_queue_open_loop_costaverage-costp2o-batch-b23ap2o-gran-per-chapterp2o-plan-bookp2o-v1queueing
In the single-server queue with service rate control, Bernoulli() arrivals, holding cost and service cost , let the allowable service rate satisfy , and let be the policy that always serves at rate . Then for every initial state the average cost of is
The quantity is the benchmark ("open loop control") against which the average cost optimal policies computed in Section 8.5 are compared.
Preamble
import Mathlib import Definitions.Def_SennottDP_AvgASM_Criteria import Definitions.Def_SennottDP_AvgASM_ServiceQueue open scoped ENNReal NNReal Topology open Filter
Formal statement
namespace SennottDP.AvgASM
/-- **Proposition 8.5.1** (Sennott 1999, p. 182). In the single-server queue with service rate
control and Bernoulli(`p`) arrivals, holding cost `H(i) = H i` and service cost `C(a)`, let the
allowable rate `a` satisfy `p < a`, and let `d(a)` be the policy that always serves at rate `a`.
Then, from every initial state `i`,
`J_{d(a)} = H p (1 − p)/(a − p) + p C(a)/a` (8.12). -/
theorem service_queue_open_loop_cost (Q : ServiceQueueData) (a : ℝ) (ha : a ∈ Q.rates)
(hpa : Q.p < a) (i : ℕ) :
avgCost (Q.serveAt a ha).toPolicy i =
ENNReal.ofReal ((Q.H : ℝ) * Q.p * (1 - Q.p) / (a - Q.p) + Q.p * (Q.serviceCost a : ℝ) / a) := by sorry
end SennottDP.AvgASM
Source
Sennott, Stochastic Dynamic Programming and the Control of Queueing Systems (Wiley, 1999), p. 182, Proposition 8.5.1, (8.12)
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.