Management Science

Behrens, Doris A.; Caulkins, Jonathan P.; Tragler, Gernot; Feichtinger, Gustav

doi: 10.1287/mnsc.46.3.333.12068pmid: N/A

Drug use and related problems change substantially over time, so it seems plausible that drug interventions should vary too. To investigate this possibility, we set up a continuous time version of the first-order difference equation model of cocaine use introduced by Everingham and Rydell (1994), extended to make initiation an endogenous function of prevalence. We then formulate and solve drug treatment and prevention spending decisions in the framework of dynamic optimal control under different assumptions about how freely drug control budgets can be manipulated. Insights include: (1) The effectiveness of prevention and treatment depend critically on the stage in the epidemic in which they are employed. Prevention is most appropriate when there are relatively few heavy users, e.g. in the beginning of an epidemic. Treatment is more effective later. (2) Hence, the optimal mix of interventions varies over time. (3) The transition period when it is optimal to use extensively both prevention and treatment is brief. (4) Total social costs increase dramatically if control is delayed.

Zhang, Z. John; Krishna, Aradhna; Dhar, Sanjay K.

doi: 10.1287/mnsc.46.3.348.12062pmid: N/A

We examine the key factors that influence a firm's decision whether to use front-loaded or rear-loaded incentives. When using price packs, direct mail coupons, FSI coupons or peel-off coupons, consumers obtain an immediate benefit upon purchase or a front-loaded incentive. However, when buying products with in-pack coupons or products affiliated with loyalty programs, promotion incentives are obtained on the next purchase occasion or later, i.e., a rear-loaded incentive. Our analysis shows that the innate choice process of consumers in a market (variety-seeking or inertia) is an important determinant of the relative impact of front-loaded and rear-loaded promotions. While in both variety-seeking and inertial markets, the sales impact and the sales on discount are higher for front-loaded promotions than for rear-loaded promotions, from a profitability perspective, rear-loaded promotions may be better than front-loaded promotions. We show that in markets with high variety-seeking it is more profitable for a firm to rear-load, and in markets with high inertia it is more profitable to front-load. Model implications are verified using two empirical studies: (a) a longitudinal experiment (simulating markets with variety-seeking consumers and inertial consumers) and (b) market data on promotion usage. The data in both studies are consistent with the model predictions.

Park, Kyungchul; Lee, Kyungsik; Park, Sungsoo; Lee, Heesang

doi: 10.1287/mnsc.46.3.363.12066pmid: N/A

We consider the node clustering problem that arises in designing a survivable two-level telecommunication network. The problem simultaneously determines an optimal partitioning of the whole network into clusters (local networks) and hub locations in each cluster. Intercluster traffic minimization is chosen as the clustering criterion to improve the service quality. Various constraints on the clustering are considered which reflect both the physical structures of local networks, such as the connectivity requirement, and the node compatibility relations such as community of interest or policy. Additional constraints may be imposed on the hub selection to ensure network survivability. We propose an integer programming formulation of the problem by decomposing the entire problem into a master problem and a number of column generation problems. The master problem is solved by column generation and the column generation problems by branch-and-cut. We develop and use strong cutting-planes for the cluster generation subproblems. Computational results using real data are reported.

Zhao, Wen; Zheng, Yu-Sheng

doi: 10.1287/mnsc.46.3.375.12063pmid: N/A

We consider a dynamic pricing model for selling a given stock of a perishable product over a finite time horizon. Customers, whose reservation price distribution changes over time, arrive according to a nonhomogeneous Poisson process. We show that at any given time, the optimal price decreases with inventory. We also identify a sufficient condition under which the optimal price decreases over time for a given inventory level. This sufficient condition requires that the willingness of a customer to pay a premium for the product does not increase over time. In addition to shedding managerial insight, these structural properties enable efficient computation of the optimal policy.Numerical studies are conducted to show the revenue impact of dynamic price policies. Price changes are set to compensate for statistical fluctuations of demand and to respond to shifts of the reservation price. For the former, our examples show that using optimal dynamic optimal policies achieves 2.4–7.3% revenue improvement over the optimal single price policy. For the latter, the revenue increase can be as high as 100%. These results explain why yield management has become so essential to fashion retailing and travel service industries.

Hazen, Gordon

doi: 10.1287/mnsc.46.3.389.12067pmid: N/A

Stochastic trees are extensions of decision trees that facilitate the modeling of temporal uncertainties. Their primary application has been to medical treatment decisions. It is often convenient to present stochastic trees in factored form, allowing loosely coupled pieces of the model to be formulated and presented separately. In this paper, we show how the notion of factoring can be extended as well to preference components of the stochastic model. We examine updateable-state utility, a flexible class of expected utility models that permit stochastic trees to be rolled back much in the manner of decision trees. We show that preference summaries for updateable-state utility can be factored out of the stochastic tree. In addition, we examine utility decompositions which can arise when factors in a stochastic tree are treated as attributes in a multiattribute utility function.

Schweitzer, Maurice E.; Cachon, Gérard P.

doi: 10.1287/mnsc.46.3.404.12070pmid: N/A

In the newsvendor problem a decision maker orders inventory before a one period selling season with stochastic demand. If too much is ordered, stock is left over at the end of the period, whereas if too little is ordered, sales are lost. The expected profit-maximizing order quantity is well known, but little is known about how managers actually make these decisions. We describe two experiments that investigate newsvendor decisions across different profit conditions. Results from these studies demonstrate that choices systematically deviate from those that maximize expected profit. Subjects order too few of high-profit products and too many of low-profit products. These results are not consistent with risk-aversion, risk-seeking preferences, Prospect Theory preferences, waste aversion, stockout aversion, or the consequences of underestimating opportunity costs. Two explanations are consistent with the data. One, subjects behave as if their utility function incorporates a preference to reduce ex-post inventory error, the absolute difference between the chosen quantity and realized demand. Two, subjects suffer from the anchoring and insufficient adjustment bias. Feedback and training did not mitigate inventory order errors. We suggest techniques to improve decision making.

Doerr, Kenneth H.; Klastorin, Theodore D.; Magazine, Michael J.

doi: 10.1287/mnsc.46.3.421.12064pmid: N/A

In this paper, we consider the design of a synchronous, unpaced flow line where workers operate at different skill levels and overtime is used, if necessary, to meet a daily production quota. The line is unpaced in the sense that items only move to the next workstation when all workers on the line have completed their respective tasks. The design problem in this case is to assign both workers and tasks to workstations to minimize the expected sum of regular and overtime costs. To solve this problem, we develop an optimization algorithm for smaller problems and a heuristic algorithm for larger problems, which we use to investigate the sensitivity of total expected cost to changes in the price of overtime, hiring practices, worker differences, and the overall amount of work time variability. Based on an extensive computational analysis, we found that (1) planned overtime is frequently beneficial, (2) more workers should be hired as worker variability increases, and (3) increases in overtime costs frequently yield a relatively lower percentage increase in total expected cost. Other managerial implications are discussed in the paper.

Chen, Frank; Drezner, Zvi; Ryan, Jennifer K.; Simchi-Levi, David

doi: 10.1287/mnsc.46.3.436.12069pmid: N/A

An important observation in supply chain management, known as the bullwhip effect, suggests that demand variability increases as one moves up a supply chain. In this paper we quantify this effect for simple, two-stage supply chains consisting of a single retailer and a single manufacturer. Our model includes two of the factors commonly assumed to cause the bullwhip effect: demand forecasting and order lead times. We extend these results to multiple-stage supply chains with and without centralized customer demand information and demonstrate that the bullwhip effect can be reduced, but not completely eliminated, by centralizing demand information.

Corbett, Charles J.; de Groote, Xavier

doi: 10.1287/mnsc.46.3.444.12065pmid: N/A

In the supply-chain literature, an increasing body of work studies how suppliers can use incentive schemes such as quantity discounts to influence buyers' ordering behaviour, thus reducing the supplier's (and the total supply chain's) costs. Various functional forms for such incentive schemes have been proposed, but a critical assumption always made is that the supplier has full information about the buyer's cost structure. We derive the optimal quantity discount policy under asymmetric information and compare it to the situation where the supplier has full information.