Informs Annual Meeting 2017

TB15

INFORMS Houston – 2017

3 - Competition for Blood Donations: A Nash Equilibrium Network Framework Pritha Dutta, Doctoral Candidate, University of Massachusetts-

3 - Scheduling Operating Rooms under the Uncertainty of Surgery Duration and the Early Cancellation Decision Guillermo Latorre-Núñez, Pontificia Universidad Católica de Chile,

Amherst, 6 H.Brandywine Drive, Amherst, MA, 01002, United States, pdutta@umass.edu, Anna B. Nagurney

Santiago, Chile, golatorre@uc.cl, Vladimir Marianov, Armin Lüer-Villagra, Jorge Vera, Germán Paredes-Belmar

In this paper, we develop a game theory model where blood service organizations compete for blood donations at collection sites in different regions based on the quality of their service and seek to maximize their transaction utilities. The governing equilibrium concept is that of Nash Equilibrium. The equilibrium is formulated as a variational inequality problem. An algorithm with nice features for computations is proposed and applied to a case study consisting of examples of increasing complexity. The results demonstrate that enhanced competition can improve the quality service level and that blood service organizations can also benefit from having collection sites in multiple regions. 4 - Column Generation Approach for Patrolling Games Abdolmajid Yolmeh, Rutgers University, Dept of Industrial & System Engineering, 96 Frelinghuysen Road, Piscataway, NJ, 08854, United States, majid_yolmeh@yahoo.com, Melike Baykal-Gursoy Most of the studies in the patrolling games area assume that all of the nodes have the same value or their values are fixed throughout time. However, in reality different nodes may have different values and these values may change over time. In this study we consider a patrolling game model on a general graph with different and dynamically changing node values and different attack times. We develop an efficient column generation approach to solve this problem. We also present a lower bound for the value of the game which can be used to terminate the column generation algorithm when a desired solution quality is reached. Computational experiments show the efficiency of the proposed column generation method. 5 - The Cooperative Game with Non-deterministic Payoffs Jian Yang, Rutgers University, 1 Washington Park Rm 1084, Newark, NJ, 07102, United States, jyang@business.rutgers.edu We study a non-traditional cooperative game where payoffs from coalitions are non-deterministic. Under various conditions, we show how the game’s core can be related to that of a traditionally defined auxiliary. When individuals both contribute to coalitions in additive fashions and share a common mean-deviation utility, we show that increased risk aversion will promote the formation of the grand coalition. Core members can also be identified by solving nonlinear or even linear programs. 332D Operations Research for Health Systems in Emerging Countries Sponsored: Manufacturing & Service Oper Mgmt, Healthcare Operations Sponsored Session Chair: Jorge Vera, Pontificia Universidad Catolica de Chile, Santiago, 7820436, Chile, jvera@ing.puc.cl 1 - A Proactive Policy to Manage Critical Inpatient Flow in Hospitals Juan-Carlos Ferrer, Pontificia Universidad Catolica de Chile, Casilla 306 Correo 22 Stgo, Santiago, Chile, jferrer@ing.puc.cl, Jaime Gonzalez Critical beds scarcity in Chilean hospitals produce overcrowding in emergency rooms, generating patients being hospitalized in urgency beds leading to large delays in emergency care. A Markov decision process was formulated to study the relation between bed availability and critical patient demand. This model shows a policy to proactively discharge patients from a unit, leaving space for an expected critical patient. The policy was tested in a simulation of a hospital in Santiago. The results show that, with the same amount of resources, both emergency patient waiting time and patient renege were reduced, increasing real capacity of the emergency room. 2 - Age Dependent Optimal Policies for Hepatitis C Treatment Susana Mondschein, Universidad Adolfo Ibanez, 2640 Diagonal las Torres, Penalolen, Santiago, 7550000, Chile, susana.mondschein@uai.cl, Natalia Yankovic Highly effective drugs for hep C virus have become available. However, their high prices require a cost effectiveness evaluation, due to the economic constraints most health care systems face. We developed a dynamic programming model with age dependent transitions, to determine optimal timing for the drug. The optimal solution is of a threshold-box type (no-yes-no, as function of the fibrosis stage). Using an age-dependent cohort, we achieved similar performance for the number of transplants, HCV deaths, and QALYs, with less overall expenditure, compared to the current Spanish policy. We also evaluated 3 heuristics, to improve both, life expectancy and expected QALYs, while containing the costs. TB14

Operating Rooms (ORs) are among the most expensive resources in a hospital. Poor scheduling of the ORs can generate delays in the schedules, excess of overtime, even cancellation of some surgeries. This directly affects the hospital costs, as well as patient and medical team satisfaction. We solve the daily OR programming problem. It consists on assigning and sequencing the surgeries to the ORs, considering uncertainty in the surgeries’ duration and a priori cancellation of surgeries. The objective is to minimize total costs. We use a robust type of approach. 4 - Admission Planning in Real Time for Emergency Services Jorge Vera, Dept. Industrial and System Engineering, Universidad Catolica de Chile, Santiago, 7820436, Chile, jvera@ing.puc.cl Jorge Vera, Pontifical Catholic University of Chile, Santiago, Chile, jvera@ing.puc.cl, Ana Celeste Batista Admission Planning Problem (APP) is concerned with deciding which patient to admit. The problem is complex due to service capacity, patient priorities and uncertainty in health care processes, such as patient length of stay. Off-line approaches have been proposed but they assume knowledge of the (uncertain) arrival rate of patients and these models would not be helpful for real-time bed allocation. We study an advance APP to control real-time bed allocation decisions for emergency patients considering on-line optimization. We derive mathematical insights to compare with the existing APP methodologies. We present computational results for a set of benchmarks in a public hospital from Chile. 332E Topics in Supply Chain Management Sponsored: Manufacturing & Service Oper Mgmt, Supply Chain Sponsored Session Chair: Nikolay Osadchiy, Emory University, Atlanta, GA, 30322, United States, nikolay.osadchiy@emory.edu 1 - The Structure of Grocery Retail Markets and Food Waste Elena Belavina, University of Chicago Booth School of Business, 5807 S.Woodlawn Ave, Chicago, IL, 60637, United States, elena.belavina@chicagobooth.edu This paper examines the impact of grocery-store density on food waste using a stylized two-echelon perishable-inventory model. We find that the effects on household and store food waste are opposite, yet household waste dominate. We find that increasing store density reduces household and total food waste, typically achieving around 6% waste reduction for a 10% reduction in the distance between stores, while making groceries more affordable—- policy makers should promote laws that encourage new store openings. Finally, popular store targeted interventions must carefully weigh the intervention’s impact on household food waste; otherwise, they might result in overall higher food waste. 2 - The Implications of Visibility on the use of Strategic Inventory in a Supply Chain Abhishek Roy, Doctoral Student, University of Texas-Austin, 2110 Speedway Stop B6500, Austin, TX, 78712-1277, United States, abhishek.roy@utexas.edu, Stephen M.Gilbert, Guoming Lai It is now widely accepted that, by holding strategic inventory, a retailer can influence the wholesale price offered by an upstream manufacturer and improve the efficiency of the supply chain. Yet, although it is typically assumed that the manufacturer can observe the retailer’s inventory, this is often not the case in practice. We recognize that this inability to observe downstream inventory creates a form of hidden action and investigate the implications upon the use of strategic inventory under both linear and non-linear contracts. 3 - Operational Transparency with Investors William Schmidt, Cornell University, 314 Sage Hall, Ithaca, NY, 14850, United States, wschmidt@cornell.edu We use two complementary approaches to tie the impact of operational disruptions to changes in the firm’s systematic risk in the aftermath of a disruption. First we show that compared to firms that comply with new transparency requirements, firms that do not comply experience a materially larger decrease in their market value after an operational disruption, even after controlling for the earnings impact of the disruption. Second, we use the change in the volatility of the firm’s share price return as a proxy for the change in the firm’s systematic risk to show that firms that do not comply experience a materially larger increase in risk after an operational disruption compared to firms that do comply. TB15

307