Capacity investment decisions under risk aversion

This article studies the optimal capacity investment problem for a risk‐averse decision maker. The capacity can be either purchased or salvaged, whereas both involve a fixed cost and a proportional cost/revenue. We incorporate risk preference and use a consumption model to capture the decision maker's risk sensitivity in a multiperiod capacity investment model. We show that, in each period, capacity and consumption decisions can be separately determined. In addition, we characterize the structure of the optimal capacity strategy. When the parameters are stationary, we present certain conditions under which the optimal capacity strategy could be easily characterized by a static two‐sided (s, S) policy, whereby, the capacity is determined only at the beginning of period one, and held constant during the entire planning horizon. It is purchased up to B when the initial capacity is below b, salvaged down to Σ when it is above σ, and remains constant otherwise. Numerical tests are presented to investigate the impact of demand volatility on the optimal capacity strategy. © 2016 Wiley Periodicals, Inc. Naval Research Logistics 63: 218–235, 2016     

Stationary (s,s) policies for a finite horizon

In the finite-horizon stochastic (s, S) inventory model with periodic review the parameters of the optimal policy generally vary with the length of the horizon. A stationary policy, however, is easier to implement and may be easier to calculate. This paper studies optimal stationary policies for a finite horizon and relates them to optimal policies through their relation to optimal stationary policies for an infinite horizon.     

A dynamic inventory system with recycling

This paper deals with a periodic review inventory system in which a constant proportion of stock issued to meet demand each period feeds back into the inventory after a fixed number of periods. Various applications of the model are discussed, including blood bank management and the control of reparable item inventories. We assume that on hand inventory is subject to proportional decay. Demands in successive periods are assumed to be independent identically distributed random variables. The functional equation defining an optimal policy is formulated and a myopic base stock approximation is developed. This myopic policy is shown to be optimal for the case where the feedback delay is equal to one period. Both cost and ordering decision comparisons for optimal and myopic policies are carried out numerically for a delay time of two periods over a wide range of input parameter values.     

Optimization of steerable sensor network for threat detection

A Markov chain approach to detecting a threat in a given surveillance zone by a network of steerable sensors is presented. The network has a finite number of predetermined states, and transition from one state to another follows a Markov chain. Under the assumption that the threat avoids detection, two game theoretic problems for finding an optimal Markov chain (two surveillance strategies) are formulated: the first maximizes the probability of threat detection for two consecutive detection periods, whereas the second minimizes the average time of detection for the worst‐case threat's trajectory. Both problems are reduced to linear programming, and special techniques are suggested to solve them. For a dynamic environment with moving noise sources, the optimal Markov chain changes at each detection period, and the rate of convergence of the Markov chain to its stationary distribution is analyzed. Both surveillance strategies are tested in numerical experiments and compared one with another. © 2011 Wiley Periodicals, Inc. Naval Research Logistics, 2011     

Optimal control of an M/G/1 queue with impatient priority customers

An optimal operating policy is characterized for the infinite‐horizon average‐cost case of a single server queueing control problem. The server may be turned on at arrival epochs or off at departure epochs. Two classes of customers, each of them arriving according to an independent Poisson processes, are considered. An arriving 1‐customer enters the system if the server is turned on upon his arrival, or if the server is on and idle. In the former case, the 1‐customer is selected for service ahead of those customers waiting in the system; otherwise he leaves the system immediately. 2‐Customers remain in the system until they complete their service requirements. Under a linear cost structure, this paper shows that a stationary optimal policy exists such that either (1) leaves the server on at all times, or (2) turns the server off when the system is empty. In the latter case, we show that the stationary optimal policy is a threshold strategy, this feature being commonplace in most of priority queueing systems and inventory models. However, the optimal policy in our model is determined by two thresholds instead of one. © 2001 John Wiley & Sons, Inc. Naval Research Logistics 48: 201–209, 2001     

Optimal pricing and inventory control policy in periodic‐review systems with fixed ordering cost and lost sales

This paper studies a periodic‐review pricing and inventory control problem for a retailer, which faces stochastic price‐sensitive demand, under quite general modeling assumptions. Any unsatisfied demand is lost, and any leftover inventory at the end of the finite selling horizon has a salvage value. The cost component for the retailer includes holding, shortage, and both variable and fixed ordering costs. The retailer's objective is to maximize its discounted expected profit over the selling horizon by dynamically deciding on the optimal pricing and replenishment policy for each period. We show that, under a mild assumption on the additive demand function, at the beginning of each period an (s,S) policy is optimal for replenishment, and the value of the optimal price depends on the inventory level after the replenishment decision has been done. Our numerical study also suggests that for a sufficiently long selling horizon, the optimal policy is almost stationary. Furthermore, the fixed ordering cost (K) plays a significant role in our modeling framework. Specifically, any increase in K results in lower s and higher S. On the other hand, the profit impact of dynamically changing the retail price, contrasted with a single fixed price throughout the selling horizon, also increases with K. We demonstrate that using the optimal policy values from a model with backordering of unmet demands as approximations in our model might result in significant profit penalty. © 2005 Wiley Periodicals, Inc. Naval Research Logistics, 2006     

Dynamic learning,pricing, and ordering by a censored newsvendor

We address the problem of determining optimal ordering and pricing policies in a finite‐horizon newsvendor model with unobservable lost sales. The demand distribution is price‐dependent and involves unknown parameters. We consider both the cases of perishable and nonperishable inventory. A very general class of demand functions is studied in this paper. We derive the optimal ordering and pricing policies as unique functions of the stocking factor (which is a linear transformation of the safety factor). An important expression is obtained for the marginal expected value of information. As a consequence, we show when lost sales are unobservable, with perishable inventory the optimal stocking factor is always at least as large as the one given by the single‐period model; however, if inventory is nonperishable, this result holds only under a strong condition. This expression also helps to explain why the optimal stocking factor of a period may not increase with the length of the problem. We compare this behavior with that of a full information model. We further examine the implications of the results to the special cases when demand uncertainty is described by additive and multiplicative models. For the additive case, we show that if demand is censored, the optimal policy is to order more as well as charge higher retail prices when compared to the policies in the single‐period model and the full information model. We also compare the optimal and myopic policies for the additive and multiplicative models. © 2007 Wiley Periodicals, Inc. Naval Research Logistics, 2007     

Optimal search using two nonconcurrent sensors

Search for a stationary target is considered for a situation in which two sensors are available, but cannot be used simultaneously. The cost (in time) of switching from one sensor to the other is ignored, and each sensor is assumed to have perfect discrimination. For a specified class of searches an optimal allocation of search effort is obtained. In the case of a circular normal prior target location distribution, an example is presented in which one sensor is assumed to have a fixed sweep width and the other a stochastic sweep width. An optimal plan is found for this example. This plan produces an allocation of search effort which is expended in a disk by one sensor and in a bounding annulus by the other.     

Optimal ordering policies for a product that perishes in two periods subject to stochastic demand

This paper considers the problem of computing optimal ordering policies for a product that has a life of exactly two periods when demand is random. Initially costs are charged against runouts (stockouts) and outdating (perishing). By charging outdating costs according to the expected amount of outdating one period into the future, a feasible one period model is constructed. The central theorem deals with the n-stage dynamic problem and demonstrates the appropriate cost functions are convex in the decision variable and also provides bounds on certain derivatives. The model is then generalized to include ordering and holding costs. The paper is concluded with a discussion of the infinite horizon problem.     

Denumerable state markovian sequential control processes: On randomizations of optimal policies

We consider a denumerable state Markovian sequential control process. It is well known that when we consider the expected total discounted income as a criterion, there exists a nonrandomized stationary policy that is optimal. It is also well known that when we consider the expected average income as a criterion, an optimal nonrandomized stationary policy exists when a certain system of equations has a solution. The problem considered here is: if there exist two optimal nonrandomized stationary policies, will a randomization of these two policies be optimal? It is shown that in the discounted case the answer is always yes, but in the average income case, the answer is yes only under certain additional conditions.     

Inventory rotation policies for slow moving items

In this article we present a stochastic model for determining inventory rotation policies for a retail firm which must stock many hundreds of distinctive items having uncertain heterogeneous sales patterns. The model develops explicit decision rules for determining (1) the length of time that an item should remain in inventory before the decision is made on whether or not to rotate the item out of inventory and (2) the minimum sales level necessary for retaining the item in inventory. Two inventory rotation policies are developed, the first of which maximizes cumulative expected sales over a finite planning horizon and the second of which maximizes cumulative expected profit. We also consider the statistical behavior of items having uncertain, discrete, and heterogeneous sales patterns using a two-period prediction methodology where period 1 is used to accumulate information on individual sales rates and this knowledge is then used, in a Bayesian context, to make sales predictions for period 2. This methodology assumes that over an arbitrary time interval sales for each item are Poisson with unknown but stationary mean sales rates and the mean sales rates are distributed gamma across all items. We also report the application of the model to a retail firm which stocks many hundreds of distinctive unframed poster art titles. The application provides some useful insights into the behavior of the model as well as some interesting aspects pertaining to the implementation of the results in a “real-world” situation.     

On the value of information in dynamic production/inventory problems under forecast evolution

In this article we explore how total system costs and inventory positions are affected when forecasts are incorporated explicitly in production/inventory systems. We assume that forecasts for demand of a certain item are available in each period, and they evolve from one period to the next in accordance with an additive evolution model. In order to analyze the effects of the forecasts on the production/inventory system we compare the optimal ordering policy and the expected costs of the model that keeps forecasts with that of a comparable standard inventory model. We show that under mild assumptions the former yields lower expected costs and inventory levels than the latter. © 1996 John Wiley & Sons, Inc.     

A multifacility capacity expansion model with joint expansion set-up costs

This article describes a multifacility capacity expansion model in which the different facility types represent different quality levels. These facility types are used to satisfy a variety of deterministic demands over a finite number of discrete time periods. Applications for the model can be found in cable sizing problems associated with the planning of communication networks. It is assumed that the cost function associated with expanding the capacity of any facility type is concave, and that a joint set-up cost is incurred in any period in which one or more facilities are expanded. The model is formulated as a network flow problem from which properties associated with optimal solutions are derived. Using these properties, we develop a dynamic programming algorithm that finds optimal solutions for problems with a few facilities, and a heuristic algorithm that finds near-optimal solutions for larger problems. Numerical examples for both algorithms are discussed.     

Fluctuations of the optimal advertising and inventory policy: A qualitative analysis

This paper discusses the properties of the inventory and advertising policy minimizing the expected discounted cost over a finite horizon in a dynamic nonstationary inventory model with random demand which is influenced by the level of promotion or goodwill. Attention is focused on the relation between the fluctuations over time of the optimal policies and the variations over time of the factors involved, i.e., demand distributions and various costs. The optimal policies are proved to be monotone in the various factors. Also, three types of fluctuations over time of the optimal policies are discussed according to which factor varies over time. For example, if over a finite interval, the random demand increases (stochastically) from one period to the next, reaches a maximum and then decreases, then the optimal inventory level will do the same. Also the period of maximum of demand never precedes that of maximum inventory. The optimal advertising behaves in the opposite way and its minimum will occur at the same time as the maximum of the inventory. The model has a linear inventory ordering cost and instantaneous delivery of stocks; many results, however, are extended to models with a convex ordering cost or a delivery time lag.     

Limiting behavior of the stochastic sequential assignment problem

The stochastic sequential assignment problem (SSAP) considers how to allocate available distinct workers to sequentially arriving tasks with stochastic parameters such that the expected total reward obtained from the sequential assignments is maximized. Implementing the optimal assignment policy for the SSAP involves calculating a new set of breakpoints upon the arrival of each task (i.e., for every time period), which is impractical for large‐scale problems. This article studies two problems that are concerned with obtaining stationary policies, which achieve the optimal expected reward per task as the number of tasks approaches infinity. The first problem considers independent and identically distributed (IID) tasks with a known distribution function, whereas in the second problem tasks are derived from r different unobservable distributions governed by an ergodic Markov chain. The convergence rate of the expected reward per task to the optimal value is also obtained for both problems. © 2013 Wiley Periodicals, Inc. Naval Research Logistics, 2013     

Multiperiod resource allocation with a smoothing objective

We consider a multiperiod resource allocation problem, where a single resource is allocated over a finite planning horizon of T periods. Resource allocated to one period can be used to satisfy demand of that period or of future periods, but backordering of demand is not allowed. The objective is to allocate the resource as smoothly as possible throughout the planning horizon. We present two models: the first assumes that the allocation decision variables are continuous, whereas the second considers only integer allocations. Applications for such models are found, for example, in subassembly production planning for complex products in a multistage production environment. Efficient algorithms are presented to find optimal allocations for these models at an effort of O(T²). Among all optimal policies for each model, these algorithms find the one that carries the least excess resources throughout the planning horizon. © 1995 John Wiley & Sons, Inc.     

A finite capacity GI/PH/1 queue with group services

We consider a finite-capacity single-server queue in which arrivals occur one at a time, according to a renewal process. The successive service times are mutually independent and have a common phase-type distribution. The customers are served in groups of size at least L, a preassigned threshold value. Explicit analytic expressions for the steady-state queue-length densities at arrivals and at arbitrary time points, and the throughput of the system are obtained. The Laplace-Stieltjes transform of the stationary waiting-time distribution of an admitted customer at points of arrivals is computed. It is shown to be of phase type when the arrival process is also of phase type. Efficient algorithmic procedures for the steady-state analysis of the model are presented. These procedures are used in arriving at an optimal value for L that minimizes the mean waiting time of an admitted customer. A conjecture on the nature of the mean waiting time is proposed.     

On the reduction of total‐cost and average‐cost MDPs to discounted MDPs

This article provides conditions under which total‐cost and average‐cost Markov decision processes (MDPs) can be reduced to discounted ones. Results are given for transient total‐cost MDPs with transition rates whose values may be greater than one, as well as for average‐cost MDPs with transition probabilities satisfying the condition that there is a state such that the expected time to reach it is uniformly bounded for all initial states and stationary policies. In particular, these reductions imply sufficient conditions for the validity of optimality equations and the existence of stationary optimal policies for MDPs with undiscounted total cost and average‐cost criteria. When the state and action sets are finite, these reductions lead to linear programming formulations and complexity estimates for MDPs under the aforementioned criteria.© 2017 Wiley Periodicals, Inc. Naval Research Logistics 66:38–56, 2019     

Multi-period airline overbooking with multiple fare classes

Consider a multi-period multi-fare class airline overbooking problem that relates to a single-leg flight. Passengers may cancel their reservations at any time, including being no-shows at flight-time. Canceling passengers receive a refund that depends on their fare class, e.g., supersaver, coach, etc. At flight-time, the airline bumps passengers in excess of flight capacity and pays a penalty for so doing. A continuous state-space dynamic programming model is developed in which the state is the numbers of reservations currently on hand in each fare class. In each period, reservation requests occur in only one fare class and the fraction of reservations canceling in each class is independent of the number of reservations therein. A booking-limit policy is optimal, i.e., in each period the airline accepts reservation requests up to a booking limit if the number of initial reservations in the fare class is less than the booking limit, and declines reservation requests otherwise. The booking limits for each class depend on the numbers of reservations in the other classes. When there are two fare classes the optimal booking limits in each class decrease with the number of reservations in the other class. © 1996 John Wiley & Sons, Inc.     

A parallel machine replacement model

This paper considers a finite horizon parallel machine replacement problem where a fixed number of machines is in operation at all times. The operating cost for a machine goes up as the machine gets older. An older machine may have to be replaced by a new one when its operating cost becomes too high. There is a fixed order cost associated with the purchase of new machines. Machine purchase prices and salvage values may depend on the period in which they were purchased. The objective is to find a replacement plan that minimizes the total discounted cost over the problem horizon. We believe that the costs in our model are more commonly observed in practice than those previously used in the literature. The paper develops properties of optimal solutions and an efficient forward‐time algorithm to find an optimal replacement plan. A dominance property is developed that further limits the options to be considered, and a simple forecast horizon result is also presented. Future research possibilities are mentioned. © 2002 Wiley Periodicals, Inc. Naval Research Logistics 49: 275–287, 2002; Published online in Wiley InterScience (http://www.interscience.wiley.com). DOI 10.1002/nav.10012