An Efficient Algorithm for Computing Optimal (s, S) Policies

  title={An Efficient Algorithm for Computing Optimal (s, S) Policies},
  author={Awi Federgruen and Paul H. Zipkin},
  journal={Oper. Res.},
This paper presents an algorithm to compute an optimal s, S policy under standard assumptions stationary data, well-behaved one-period costs, discrete demand, full backlogging, and the average-cost criterion. The method is iterative, starting with an arbitrary, given s, S policy and converging to an optimal policy in a finite number of iterations. Any of the available approximations can thus be used as an initial solution. Each iteration requires only modest computations. Also, a lower bound on… 

Tables from this paper

Finding Optimal (s, S) Policies Is About As Simple As Evaluating a Single Policy

In this paper, a new algorithm for computing optimal ( s , S ) policies is derived based upon a number of new properties of the infinite horizon cost function c ( s , S ) as well as a new upper bound

Computing optimal (s, S) policies in inventory models with continuous demands

An iterative algorithm for continuous demand distributions which avoids any form of prior discretization is presented which can be viewed as a modified form of policy iteration applied to a Markov decision process with continuous state space.

On asymptotic approximations for (s,S) policies

This paper is concerned with optimal policy approximations based on asymptotic renewal theory, and their accuracy conditions, for a class of periodic-review inventory systems We compare the

A simple proof for optimality of (s, S) policies in infinite-horizon inventory systems

The classical proofs for the existence of a stationary (s, S) inventory policy that minimizes the total discounted or average cost over an infinite horizon are lengthy because they depend heavily on

A recursion-free functional approximation for the dynamic inventory problem

A recursion-free approximation for the optimal cost function of the dynamic inventory problem with non-stationary demands is devised, which enables us to compute policy parameters heuristically, without resorting to a stochastic dynamic program.

Error Analysis of an Approximate Optimal Policy for an inventory System with Stochastic and continuous Demands

This paper provides an approach, by which an analytical bound of the error between the costs of the approximate optimal policy and the optimal policy is given, and it is shown that the approximate ideal policy converges to the optimalpolicy as the size of the computational grid decreases.

A new algorithm for computing optimal (s, S) policies in a stochastic single item/location inventory system

A new algorithm to search for the optimal pair of s and S is proposed, which differs from the approach of Zheng and Federgruen (1991) in several aspects and has certain advantages.

Control-limit policies for a class of stopping time problems with termination restrictions

We consider a class of stopping time problems in which the state of the process evolves according to a discrete-time Markov chain and the action which terminates the process is available only in a



Convergence results and approximations for optimal (s,s) policies

In this paper we consider the dynamic inventory model with a discrete demand and no discounting. We verify a conjecture of Iglehart about the asymptotic behaviour of the minimal total expected cost.

An Empirical Study of Exactly and Approximately Optimal Inventory Policies

The (s, S) inventory model (infinite horizon with stationary parameters) is explored with a view toward determining the dependency of the optimal policy on the model's parameters and toward

The Power Approximation for Computing (s, S) Inventory Policies

A new analytic approximation for computing (s, S) policies for single items under periodic review with a set-up cost, linear holding and shortage costs, fixed replenishment lead time, and backlogging of unfilled demand is presented.

A general markov decision method I: Model and techniques

In the policy-iteration algorithm resulting from this approach the number of equations to be solved in any iteration step can be substantially reduced, and by its flexibility, this algorithm allows us to exploit any structure of the particular problem to be solve.

A Method for Approximate Solutions to Stochastic Dynamic Programming Problems Using Expectations

The technique is to replace probability distributions by their corresponding expectations, and to use the values of the states in the corresponding deterministic system under its optimal policy to determine an approximate policy in the stochastic system through a single application of Howard's policy improvement operation.

(s, S) Policies for a Dynamic Inventory Model with Stochastic Lead Times

This study analyzes a stochastic lead time inventory model under the assumptions that a replenishment orders do not cross in time and the lead time distribution for a given order is independent of the number and sizes of outstanding orders to yield an intuitively appealing dynamic program.

Improved Algorithms for Inventory and Replacement-Stocking Problems

This paper is devoted to improving known schemes for computing optimal policies in stochastic decision problems. Improved procedures are developed for computing optimal $( s,S)$ policies in the

Evaluating the Effectiveness of a New Method for Computing Approximately Optimal (s, S) Inventory Policies

A new method for obtaining approximately optimal (s, S) inventory policies was presented recently as an illustration of a general approximation methodology proposed by Porteus and numerical work showed that when demand was normally distributed the method performed very well.

Computing Optimal Control Limits for GI/M/S Queuing Systems with Controlled Arrivals

This paper shows how the special structure of this queuing system can be exploited to develop efficient procedures to determine an optimal control limit and presents an explicit algorithm for computing the optimal control.

s, S Policies Under Continuous Review and Discrete Compound Poisson Demand

We consider inventory policies for a continuous review system with discrete compound Poisson demand, convex holding-penalty cost and fixed order cost. Beckmann has showri that under these conditions