We consider congestion control in a non-stationary queueing system. Assuming that the arrival and service rates are measurable, bounded, and periodic functions of time, a Markov
decision process (MDP) formulation is developed. We show under infinite horizon discounted and average reward optimality criterion, for each fixed time, optimal pricing and admissioncontrol strategies are non-decreasing in the number of customers in the system. This extends
well-known stationary results to the non-stationary setting. However, even with this insight,
solution of the problem is intractable. We propose an easily implementable
pointwise stationaryapproximation
(PSA) to approximate the optimal policies in each case, and suggest a heuristicto improve the implementation of the PSA and verify its usefulness via a numerical study.