2016 INFORMS Annual Meeting Program

TB17

INFORMS Nashville – 2016

3 - Product Planning With Sensory Customer Requirements: A Plain Yogurt Case Leman Esra Dolgun, Anadolu University, Iki Eylÿl Kampusu,

that each patient may have different priority. Our model considers patient segments with different priorities to design an efficient yet profitable scheduling system. 2 - Mutual Information Minimization For Evaluating The Causal Impact Of Home Care Services On Patient Discharge Disposition Alexander Nikolaev, University at Buffalo, anikolae@buffalo.edu, Yuan Zhou, Sabrina Casucci, Lei Sun, Li Lin Recent developments in observational causal inference employ subset selection algorithms that balance the covariate distributions in the compared groups of treated and untreated units under study. We present a subset selection approach that works by minimizing the mutual information (MI) between the covariates and the treatment variable. This becomes possible thanks to the derived optimality conditions that tackle the non-linearity of a sample-based MI function. The resulting algorithm runs in polynomial (close to linear) time, allowing for treatment effect estimation with large datasets. We proceed to draw causal insights from hospital readmission data of the UB HOMEBASE project. 3 - Information Acquisition Process: The Quantitative Aspect Eugene Perevalov, Lehigh University, eup2@lehigh.edu The classical Information Theory was able to facilitate significant advances in information transmission by finding a way of properly describing the quantity of the form of information existence - the various symbols. On the other hand, to make similar advances in the process of information acquisition, one would need to properly understand and describe the quantities associated with the information itself - its content. We take a step in that direction. Stochastic Sequential Assignment and Planning Sponsored: Optimization, Optimization Under Uncertainty Sponsored Session Chair: Olga Raskina, Juno Therapeutics, Seattle, WA, United States, olga.raskina@junotherapeutics.com 1 - Stochastic Production Scheduling For Personalized Cancer Therapy Manufacturing Olga Raskina, Juno Therapeutics, olga.raskina@junotherapeutics.com, Jon Gunther Juno Therapeutics is a clinical-stage company developing novel cellular immunotherapies. To manufacture our therapeutic T cell product candidates, we harvest blood cells from a cancer patient, separate or enrich for the appropriate T cells, activate the cell, insert the gene sequence for the CAR or TCR construct into the cell’s DNA, and grow these modified T cells to the desired dose level. The modified T cells can then be infused into the patient or frozen and stored for later infusion. We are investing substantially in a process that we believe is commercially scalable for both CARs and TCRs. 2 - Dynamic Path Assignment For Robotic Bees Under Collision Control Felisa Vázquez-Abad, Hunter College, felisav@hunter.cuny.edu We study the problem of dynamic routing of robotic bees towards the hive. Due to uncertainty in position measurements, the stochastic problem cannot ensure collision-free paths. We study the effects that the algorithm parameters have in reducing the computational complexity and expected number of collisions. We explore via experimentation a parallel algorithm for cloud computing as well as crowd sourcing for estimation. In this manner we seek a faster completion time reducing collisions and computational complexity. 3 - Multidimensional Birth And Death Models For Biochemical Modeling With A View To Personalized Medicine Alexey Nikolaev, CUNY, Hunter College and Graduate Center, New York, NY, 11232, United States, anikolaev@gradcenter.cuny.edu, Felisa J Vázquez-Abad, Jon Gunther Chemical reactions can be modeled either using ordinary differential equations, or continuous time Markov chains. The ODE describes the dynamics of the concentration of proteins. The multidimensional Birth and Death process describes the dynamics of the number of particles of proteins. We study an activator/repressor model that regulates circadian cycles in cells. We show that the ODE is an exact expectation for the B&D process, under some regimes. For other regimes we apply a novel technique of Taboo kernels. This methodology has potential to predict malfunctions and to identify treatment. TB17 105B-MCC

Endustri Muhendisligi Bolumu, Eskisehir, Turkey, ledolgun@gmail.com, Gulser Koksal, Elcin Kartal Koc

Customers use vague terms to describe their requirements and perceptions of sensory products such as yogurt, which makes prioritizing improvements and setting targets for them difficult. In this study, we propose an improved use of Quality Function Deployment and Kansei Engineering to overcome such difficulties, and demonstrate it in a plain yogurt case. Customer needs and perceptions are assessed by special surveys based on semantic scales. Survey data are analyzed using hypothesis tests and building logistic regression models. Targets are suggested by optimizing these models. 4 - A Reformulation-based Interior-point Path-following Method For Determining Proper Equilibria Chuangyin Dang, Professor, City University of Hong Kong, Dept of Systems Eng & Eng Mgmt, 83 Tat Chee Avenue, Kowloon, Hong Kong, mecdang@cityu.edu.hk, Yin Chen This paper extends an equivalent reformulation of Nash equilibrium to a specifically defined perturbed game that deforms continuously with an extra variable. As a result of this extension, we develop an interior-point path-following method for determining a proper equilibrium. The method numerically follows a smooth path that starts from any given point in the Euclidean space and ends at a proper equilibrium. Numerical results further confirm the effectiveness of the method. Industry Job Search Panel Invited: INFORMS Career Center Invited Session Moderator: Adam McElhinney, Uptake Technologies, Chicago, IL, United States, adam.m.mcelhinney@gmail.com 1 - Industry Job Search Panel Adam McElhinney, Uptake Technologies, Chicago, IL, United States, adam.m.mcelhinney@gmail.com Ever wonder how to make the transition from student to OR, Analytics or Data Science professional? Or maybe you are an experienced professional looking to make a career switch. Join the Industry Job Search Panel to hear tips from the pros. Specifically: 1. Strategies to ensure your skills are aligned with the job market 2. Tips and tricks for landing and succeeding at the interview 3. How to interview the interviewer 2 - Panelist Aly Megahed, IBM Research - Almaden, 150 Palm Valley Blvd APT 2066, San Jose, CA, 95123, United States, aly.megahed@us.ibm.com 3 - Panelist Beverly Wright, Cox Communications Inc, 125 Centennial Drive, Peachtree City, GA, 30269, United States, Beverly.Wright@scheller.gatech.edu 4 - Panelist Warren Hearnes, Cardlytics, Lilburn, GA, 30047, United States, whearnes@hotmail.com TB16 105A-MCC Information in Optimization Sponsored: Optimization, Optimization Under Uncertainty Sponsored Session Chair: Eugene Perevalov, Lehigh University, 1, Bethlehem, PA, 18015, United States, eup2@lehigh.edu 1 - Outpatient Clinic Scheduling With Heterogeneous Patient Preference And Resource Uncertainty Deepak Agrawal, Pennsylvania State University, University Park, PA, United States, agrawal.deepankur@gmail.com, Guodong Pang, Priyantha Devapriya, Soundar Kumara Reasons of No-shows and how to stop them, has been focus of the research since more than a decade. No-show adds up to the increases healthcare waste. Therefore motivated by this we aim to develop advanced scheduling models which can reduce No-shows by scheduling appointments at patients’ preferred day and time with their preferred physicians. Researchers have acknowledged TB15 104E-MCC

269