2016 INFORMS Annual Meeting Program

SD79

INFORMS Nashville – 2016

5 - Probability Analysis Of The Severity Of Train Derailments Using Copula Models Emmanuel Martey, University of Delaware, 302 DuPont Hall, Newark, DE, 19716, United States, enmartey@udel.edu, Nii Attoh-Okine In spite of their relatively low occurrence, train derailments have been a major concern due to their high consequence. Derailment severity may depend on various factors such as speed, accident cause and residual train length. It is important to know the dependencies between these variables in order to better understand how to reduce derailment severity. This paper presents the copula approach as a technique for modeling dependencies between the various variables. Copulas link arbitrary marginal distributions to form a joint multivariate distribution with a particular dependence structure. Copulas are suitable for modelling multivariate data with non-normality, tail dependency or skewness. SD72 Bass- Omni Supply Chain Mgt IV Contributed Session Chair: Shabnam Rezapour, University of Oklahoma, 2248 Houston Ave. apt 2, Norman, OK, 73071, United States, shabnam_rezapoor@yahoo.com 1 - Upstream Supplier And Downstream Customer Networks: An Empirical Investigation Marcus A Bellamy, Boston University, Rafik B. Hariri Building, 595 Commonwealth Avenue, Boston, MA, 02215, United States, bellamym@bu.edu, Soumen Ghosh, Manpreet Singh Hora We examine the relationship dependence characteristics and structural configuration of a firm’s supply chain as drivers of its performance using supply chain relationship data from the Bloomberg database. We demonstrate how firm performance may be influenced by the manner in which its cost is concentrated upstream as a customer, its revenue is concentrated downstream as a supplier, and its supply network is structured. 2 - Capacity Expansion Under Demand Uncertainty With Uncertain Probabilities Heejung Kim, University of California- Berkeley, Berkeley, CA, 94720-1777, United States, kimheejung@berkeley.edu, Philip Kaminsky Pharmaceutical industries make capacity investment decisions while clinical trials for products are running. The demands are highly dependent on the test results, and estimating exact probability distribution of the results is difficult. We focus on developing and understanding capacity expansion models that are robust to any possible probability distributions using multistage stochastic programming for different objectives - minimizing expected cost, value at risk and conditional value at risk. 3 - Supply Chain Partner Environmental Health And Firm Performance Marcus A Bellamy, Assistant Professor, Boston University, Rafik B. Hariri Building, 595 Commonwealth Avenue, Boston, MA, 02215, United States, bellamym@bu.edu We empirically examine the relationship between the environmental initiatives and outcomes of a firm’s supply chain partners and firm performance. We draw from environmental, financial, and supply chain data to identify key mechanisms related to the environmental health of a firm’s supply chain that influence its economic performance. 4 - Component Procurement And End Product Assembly In An Uncertain Supply And Demand Environment Ramesh Bollapragada, San Francisco State University, School of Business, 1600 Holloway Avenue, San Francisco, CA, 94132, United States, rameshb@sfsu.edu, Saravanan Kuppusamy, Uday S Rao In this paper, we examine a multi-product, multi-component, procurement and assembly problem with both supply and demand uncertainties. We explicitly model the uncertainty using a stochastic program that facilitates procurement and assembly decisions. We present a stochastic linear programming model of the problem which we solve using its deterministic equivalent with a finite number of scenarios. We identify the key cost drivers that need attention from managers in the manufacturing industry, when there is limited knowledge of future demand and component availability.

5 - Correlation Between Supply Networks’ Strategic And Operational Risk Mitigation Strategies Shabnam Rezapour, University of Oklahoma, 2248 Houston Ave. apt 2, Norman, OK, 73071, United States, shabnam_rezapoor@yahoo.com, Janet K. Allen, Farrokh Mistree A supply network’s performance is affected by two types of uncertainty: 1) disruptions distorting its topology; and 2) variations affecting its flow planning. We show that strategic risk mitigation strategies, such as robustness and resilience, and operational risk mitigation strategies, such as reliability, neutralizing impacts of disruptions and variations respectively are correlated. A model is developed to simultaneously determine robustness, resilience and reliability. Our findings show that the correlation between: i) robustness and resilience is negative; ii) robustness and reliability is positive; and iii) resilience and reliability is negative.

SD79 Legends G- Omni Health Care, Modeling IV Contributed Session

Chair: Utpal Kumar Bhattacharya, Associate Professor, Indian Institute of Management Indore, Pitampur Road, Prabandh Sikhar, Indore, 453556, India, utpalb@iimidr.ac.in 1 - Optimal Radiotherapy Treatment Policy Based On Tumor Biological Response: A Partially Observable Markov Decision Process Framework Nasrin Nouri, PhD Student, University of Houston, 9701 Meyer Forest Dr., Apt 6207, Houston, TX, 77096, United States, nouri.nasrin@gmail.com In radiotherapy treatment planning the prescribed dose is delivered in equal fractions of dose during 30 to 40 sessions to give healthy organs time to recover. Depending on tumor state, the tumor growth and its response to radiation will change, hence a dynamic treatment plan is required. It is not possible to observe the tumor before each session through CT images so we are faced to uncertainty of tumor state. In this study we develop a partially observable Markov decision process to provide optimal treatment policy when the density of tumor is uncertain. This approach provides the optimal policy determining when to choose a less effective, less harmful dose over a more effective, more harmful dose. 2 - Reserving Walk-in Times In Primary Care For a primary care physician with varying workday demand, capacity reservation for walk-ins and scheduled appointment slots is optimized on a tactical level. Number and position of the scheduled appointments influence waiting times for patients, capacity for treatment and the utilization of PCPs. A multi-criteria mixed-integer linear programming model is suggested to find an acceptable compromise solution. Results are evaluated by an extensive stochastic simulation study. 3 - Econometric Model Of Critical Care Outreach Team And Intensive Care Unit Ali Haji Vahabzadeh, The University of Auckland Business School, Private Bag 92019, Auckland, 1142, New Zealand, a.vahabzadeh@auckland.ac.nz, Valery Pavlov To analyse the role and functionality of the critical care outreach team (CCOT) in hospitals, and particularly, its interactions with the ICU, we develop an econometric model of CCOT and ICU. This allows us to estimate the impact of CCOT intervention in detecting the critically ill patients in the ward on the ICU length-of-stay, potential ICU readmission and patient outcome. Brigitte Werners, Professor, Ruhr-University Bochum, Fac. Management and Economics, Bochum, 44780, Germany, or@rub.de

119