Search

The interest in applying machine learning in healthcare has grown rapidly in recent years. Most predictive algorithms requiring pathway implementations are evaluated using metrics focused on predictive performance, such as the c statistic. However, these metrics are of limited clinical value, for two reasons: (1) they do not account for the algorithms role within a provider workflow; and (2) they do not quantify the algorithms value in terms of patient outcomes and cost savings. We propose a model for simulating the selection of patients over time by a clinician using a machine learning algorithm, and quantifying the expected patient outcomes and cost savings. Using data on unplanned emergency department surgical readmissions, we show that factors such as the providers schedule and postoperative prediction timing can have major effects on the pathway cohort size and potential cost reductions from preventing hospital readmissions.

When suppliers (i.e., contract manufacturers) fail to comply with health and safety regulations, buyers (retailers) are compelled to improve supplier compliance by conducting audits and imposing penalties. As a benchmark, we first consider the independent audit-penalty mechanism in which the buyers conduct their respective audits and impose penalties independently. We then examine the implications of two new audit-penalty mechanisms that entail a collective penalty. The first is the joint mechanism under which buyers conduct audits jointly, share the total audit cost incurred, and impose a collective penalty if the supplier fails their joint audit. The second is the shared mechanism in which each buyer conducts audits independently, shares its audit reports with the other buyers, and imposes a collective penalty if the supplier fails any one of the audits. Using a simultaneous-move game-theoretic model with two buyers and one supplier, our analysis reveals that both the joint and the shared mechanisms are beneficial in several ways. First, when the wholesale price is exogenously given, we establish the following analytical results for the joint mechanism in comparison with the independent mechanism: (a) the supplier’s compliance level is higher; (b) the supplier’s profit is lower while the buyers’ profits are higher; and (c) when the buyers’ damage cost is high, the joint audit mechanism creates supply chain value so the buyers can offer an appropriate transfer payment to make the supplier better off. Second, for the shared audit mechanism, we establish similar results but under more restrictive conditions. Finally, when the wholesale price is endogenously determined by the buyers, our numerical analysis shows that the key results continue to hold. The online appendix is available at https://doi.org/10.1287/msom.2017.0653 .

We analyzed the planning problem for HIV screening, testing, and care. This problem consists of determining the optimal fraction of patients to be screened in every period as well as the optimum staffing level at each part of the healthcare system to maximize the total health benefits to the patients measured by quality-adjusted life-years (QALYs) gained. We modeled this problem as a nonlinear mixed integer programming program comprising disease progression (the transition of the patients across health states), system dynamics (the flow of patients in different health states across various parts of the healthcare delivery system), and budgetary and capacity constraints. We applied the model to the Greater Los Angeles (GLA) station in the Veterans Health Administration system. We found that a Centers for Disease Control and Prevention recommended routine screening policy in which all patients visiting the system are screened for HIV irrespective of risk factors may not be feasible because of budgetary constraints. Consequently, we used the model to develop and evaluate managerially relevant policies within existent capacity and budgetary constraints to improve upon the current risk based screening policy of screening only high risk patients. Our computational analysis showed that the GLA station can achieve substantial increase (20% to 300%) in the QALYs gained by using these policies over risk based screening. The GLA station has already adapted two of these policies that could yield better patient health outcomes over the next few years. In addition, our model insights have influenced the decision making process at this station.

In this paper we apply emerging theories in probability and statistics to examine the value of pooling of inventories under arbitrary non-normal dependent demand structures. Eppen (1979) showed that inventory costs in a centralized system increase with the correlation between multivariate normal product demands; using multivariate stochastic orders, we generalize this statement to arbitrary distributions. We then describe methods to construct models with arbitrary dependence structure, using the copula of a multivariate distribution to capture the dependence between the components of a random vector. For broad classes of distributions with arbitrary marginals, we confirm that pooling of inventories is more valuable when demands are less positively dependent.

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.ejor.2006.06.054 Byline: Rui Yin, Kumar Rajaram Keywords: Joint pricing and inventory control; Markovian demand; Optimal feedback policy Abstract: We consider the joint pricing and inventory control problem for a single product over a finite horizon and with periodic review. The demand distribution in each period is determined by an exogenous Markov chain. Pricing and ordering decisions are made at the beginning of each period and all shortages are backlogged. The surplus costs as well as fixed and variable costs are state dependent. We show the existence of an optimal (s, S, p)-type feedback policy for the additive demand model. We extend the model to the case of emergency orders. We compute the optimal policy for a class of Markovian demand and illustrate the benefits of dynamic pricing over fixed pricing through numerical examples. The results indicate that it is more beneficial to implement dynamic pricing in a Markovian demand environment with a high fixed ordering cost or with high demand variability. Author Affiliation: UCLA Anderson School of Management, The University of California, P.O. Box 951481, 110 Westwood Plaza, Los Angeles, CA 90095-1481, United States Article History: Received 9 November 2005; Accepted 24 June 2006

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.ejor.2006.03.020 Byline: Inneke Van Nieuwenhuyse (a), Nico Vandaele (b), Kumar Rajaram (c), Uday S. Karmarkar (c) Keywords: Queueing; Campaign sizing; Product allocation; (Semi)process industries Abstract: This paper studies the impact of management policies, such as product allocation and campaign sizing, on the required size of the finished goods inventories in a multi-product multi-reactor batch process. Demand, setup and batch processing times for these products are assumed to be stochastic, and the inventory buffer for every product type needs to be such that target customer service levels are met. To perform this analysis, we develop a queueing model that allows us to explicitly estimate service levels as a function of the buffer size, and the allocation/campaign sizing policies. This model can be used to evaluate the service level given an existing buffer configuration, as well as to determine the buffer sizes required across products to meet a pre-specified service level. It also allows us to formulate a number of insights into how product allocation decisions and campaign planning policies affect buffer sizing decisions in symmetric production systems. Author Affiliation: (a) European University College Brussels, Stormstraat 2, 1000 Brussels, Belgium (b) Faculty of Applied Economics, University of Antwerp, Prinsstraat 13, 2000 Antwerp, Belgium (c) Decision, Operations and Technology Management, UCLA Anderson School of Management, Box 951481, Los Angeles, CA 90095-1481, United States Article History: Received 4 March 2005; Accepted 16 March 2006

The effects of management policies on finished goods inventory size, using queueing model, are presented.

To link to full-text access for this article, visit this link: http://dx.doi.org/10.1016/j.ejor.2005.11.009 Byline: Kevin F. McCardle, Kumar Rajaram, Christopher S. Tang Keywords: Retailing; Product bundling; Pricing; Inventory management; Basic products; Fashion products Abstract: We consider the impact of bundling products on retail merchandising. We consider two broad classes of retail products: basic and fashion. For these product classes, we develop models to calculate the optimal bundle prices, order quantities, and profits under bundling. We use this analysis to establish conditions and insights under which bundling is profitable. Our analysis confirms that bundling profitability depends on individual product demands, bundling costs, and the nature of the relationship between the demands of the products to be bundled. We also provide detailed numerical examples. Author Affiliation: University of California at Los Angeles, Anderson Graduate School of Management, 110 Westwood Plaza, Box 951481, Los Angeles, CA 90095-1481, United States Article History: Received 2 August 2004; Accepted 27 November 2005 Article Note: (footnote) [star] The authors thank Professor Steven Lippman and three anonymous referees for their comments and suggestions.

Eppen (1979) showed that inventory costs in a centralized system increase with the correlation between multivariate normal product demands. Using multivariate stochastic orders, we generalize this statement to arbitrary distributions. [...]

The impact of variability on a continuous flow production process is analyzed. A continuous time model is developed to capture the propagation of variability through the system and use this model to calculate the mean and the variance of the distribution of the output from this process.

We analyze planning and scheduling of multiproduct batch operations in the food-processing industry. Such operations are encountered in many applications including manufacturing of sorbitol, modified starches, and specialty sugars. Unlike [...]

An interactive decision support system for operational decision-making at a starch factory is presented. The technique discussed is for online process control.

We consider the problem of determining (for a short lifecycle) retail product initial and replenishment order quantities that minimize the cost of lost sales, back orders, and obsolete inventory. We [...]

We consider the anesthesiologist staff planning problem for operating services departments in large multi‐specialty hospitals without limit on anesthesiologist supply, where the planner makes monthly and daily decisions to minimize total costs. Each month the staff planner decides the number of anesthesiologists on regular duty and an on‐call consideration list for each day of the following month. In addition, each day, the staff planner decides how many on‐call anesthesiologists to call for the following day. Total costs consist of explicit and implicit costs. Explicit costs include the costs of calling an anesthesiologist and overtime costs. These costs are specified by the organization. Implicit costs encompass costs of not calling an on‐call anesthesiologist and under‐utilizing an anesthesiologist, and these have to be deduced from past decisions. We model the staff planning problem as a two‐stage integer stochastic dynamic program. We develop structural properties of this model and use them in a sample average approximation algorithm constructed to solve this problem. We also develop a procedure to estimate the implicit costs, which are included in this model. Using data from the operating services department at the UCLA Ronald Reagan Medical Center, our model shows the potential to reduce overall costs by 16%. We provide managerial insights related to the relative scale of these costs, hiring decisions by service, sensitivity to cost parameters, and improvements in the prediction of the booked time durations. [ABSTRACT FROM AUTHOR]

An investigation is conducted into the impact of product substitution on retail merchandising, with particular focus on two aspects, anticipated profits and order quantities. The basic news-vendor model is extended as a method of performing this evaluation.

The use of assortment planning in fashion retailing is examined.

We consider the distribution planning problem in the motion picture industry. This problem involves forecasting theater level box office revenues for a given movie and using these forecasts to choose the best locations to screen a movie. We first develop a method that predicts theater-level box office revenues over time for a given movie as a function of movie attributes and theater characteristics. These estimates are then used by the distributor to choose where to screen the movie. The distributor’s location selection problem is modeled as an integer programming-based optimization model that chooses the location of theaters in order to optimize profits. We tested our methods on realistic box office data and show that it has the potential to significantly improve the distributor’s profits. We also develop some insights into why our methods outperform existing practice, which are crucial to their successful practical implementation. [ABSTRACT FROM AUTHOR]