Informs Annual Meeting 2017

POSTER SESSION

INFORMS Houston – 2017

8 - Best Subset Selection of Factors Affecting Influenza Spread using Multi Objective Optimization Aigerim Bogyrbayeva, graduate Student, University of South Florida, 4202 E. Fowler Avenue, Tampa, FL, 33620, United States, aigerimb@mail.usf.edu, Hanisha Tatapudi, Walter Silva, Hadi Charkhgard Influenza ranks among the most common viral infections. It is extremely contagious and fatal for senior adults and children. Influenza is treated by pharmaceutical interventions (i.e. vaccines) and non-pharmaceutical interventions (i.e. quarantines). Due to the nature of the disease and its viral mutations, creating vaccines for every Influenza strain is time consuming. Hence, Non-Pharmaceutical Interventions (NPIs) have been successfully used in the past. We study the impact of NPIs on the spread of the disease and recommend the best subset selection of NPIs. We used a bi-objective optimization approach for accurately predicting the infection attack rate and minimizing the number of NPIs. 9 - Will the High-speed Passenger Railway Line Operation Affects on Regional Logistics System Mi Gan, Southwest Jiaotong University, Jinniu District, 111 N.1st Erhuan Road, Chengdu, 610031, China, migan@swjtu.cn, Xi Huang An applied STIRPAT model has been applied to estimate the impacts of high speed passenger railway line operation on regional logistics system. The regional freight volume from 2014-2016 has been selected to represent the status of logistics system, the number of passenger of each stations of high speed passenger railway line has been selected as the independent variable. The analysis result shows high-speed rail operation has positive effects on regional logistics system. 10 - An Optimization Model for Resilient Service Systems Jing Gong, Jinan University, School of Public Administration, Guangzhou, China, tgongjing@jnu.edu.cn A service system in our research has two-fold meaning: 1)to meet the demand; 2)to be delivered by a directed graph. Recent catastrophic events precipitate a wide range of impacts so that public awareness continues to grow of resilience as an important means of characterizing the ability to deal with them. Taking resilience into account will raise the following questions: How can we represent the resilience of a service system? How can managers tradeoff between a cost minimum system and a resilience maximum system? This paper proposes a linear piecewise recovery function and applies it to the decision-making processes of improving a service system’s resilience to disasters. 11 - Educational Efficiency Distance in Secondary Schools of SãO Paulo Municipalities Maria Cristina Gramani, Professor, INSPER.Institute of Education and Research, Rua Quatá 300, Sao Paulo, 04546-042, Brazil, mariacng@insper.edu.br This study analyzes the performance of secondary schools in Brazil according to two indicators: (i) “gap of the school”, which measures the real distance of performance between schools in national ranking and, (ii) “discipline bias”, which investigates whether schools with better educational performance are those with a higher tendency to score better in certain fields. We found that, high level of students in Mathematics consists on the determinant of efficiency of schools with a higher socioeconomic level, and the lower number of students with low level in Human science is the efficiency determinant of schools with lower socioeconomic status. 12 - Determining Decoupling Points in a Supply Chain Associated with Different Design Preferences Lin Guo, University of Oklahoma, Norman, OK, United States, lin.guo@ou.edu, reza alizadeh, Janet Katherine Allen, Farrokh Mistree In this research, we propose an approach using exploration of the solution space of a multi-goal compromise Decision Support Problems model to determine the decoupling points of a three-echelon supply chain. Facing different design preferences of the supply chain, the goals should be satisfied in different degree, and the change of the design preference may take place rapidly, which requires the decoupling points flexible enough to be moved through the supply chain. 13 - Bus Ranking Method for Robust Optimization Amandeep Gupta, Cornell University, 316 Highland Road, Ithaca, NY, 14850, United States, ag729@cornell.edu, Lindsay Anderson Interest in efficient and flexible stochastic methods for unit commitment has been growing steadily in answer to increasing renewable energy resources on the grid. This work describes a statistical bus ranking methodology that identifies the most critical buses based on criteria such as economic dispatch cost, or ramping needs to provide a robust unit commitment solution. The method is illustrated via a case study on the IEEE 30-bus system and compared to other established approaches to demonstrate the efficacy of obtained solutions. Results show that the bus ranking method performs as well as the best of these methods, with the provision of additional flexibility and potential for parallelization.

14 - Optimal Step-stress ALT under Interval Monitoring and Progressive Type-I Censoring David Han, University of Texas at San Antonio, Management Science & Statistics, College of Business, San Antonio, TX, 78249- 0632, United States, david.han@utsa.edu A step-stress ALT under progressive Type-I censoring was considered when a continuous monitoring of failures is infeasible but inspections at certain time points is possible. In addition to the accelerated failure time model, a general scale family of distributions was considered for flexible modeling. The MLE of scale parameters and their conditional densities could be derived explicitly. Using the exact distributions of the estimators, confidence intervals for the parameters were obtained and numerically assessed. 15 - Views and Subscriptions: Which One Matters to a Youtuber’s Monetization Success? The critical factors to a monetization success have gained a tremendous interest from YouTubers and social media entrepreneurs. However, there has not been a rigorous study testing the effects of the view factors and subscription factors on a YouTuber’s revenues. We introduce a new empirical model to address this knowledge gap. Validating our model by a panel data of 116 most viewed YouTuber channels, we find that the new daily views, the daily view growth rate, and the existing total views of a YouTuber’s videos are significantly positively associated with her daily revenue. The subscription variables do not directly contribute to a YouTuber’s daily revenue. 16 - Server Behaviours in Queueing Systems Paul Harper, Cardiff University, School of Mathematics, Cardiff, CF24 4AG, United Kingdom, harper@cardiff.ac.uk In the classical queueing theory literature, a server is commonly assumed to work at a constant speed. Motivated by observations from healthcare applications, in particular within Emergency Departments, a study is made to explore the nature of the relationship between service times and workload in order to assess and quantify any workforce (server) behaviours. An analytical model is presented in which service time depends on queue length. The research helps to demonstrate the importance of more accurately capturing server behaviours in workload- dependent environments, and the impact this has on the overall system performance. 17 - Optimization of Emergency Evacuations Through Exit Allocation and Risk Analysis Elizabeth Hooper, Texas A&M.University-Kingsville, Kingsville, TX, United States, elizabeth@alumni.purdue.edu Exit locations and evacuation procedures facilitate occupant safety during emergencies. This research aims to optimize building exit locations and analyze risk during building evacuations. Nonlinear programming and queueing theory are used to optimize exit locations and determine egress times. A risk analysis with risk matrix and flow chart for emergency evacuation are also created. Lastly, this paper discusses an application of smart building technology for building evacuations during emergencies. The results show that the use of optimization techniques and risk analysis allow better planning and response during emergencies. 18 - Decision Pattern Mining in Chronic Disease Praowpan Tansitpong, Mahidol University International College, 999 Bhuddamonthn 4 Rd., Nakon Pathom, 73170, Thailand, praowpan.tan@mahidol.ac.th This study utilizes electronic medical records (EMRs) to evaluate diversity in standard of care from prescription decision in multiple clusters and identify “high cost” segment by using health benefit differentiation, brick-and-mortar Pareto classification, and Hierarchical clustering. 19 - Natural Language Processing Techniques for Patent Analysis in Management Research Mike Horia Teodorescu, Doctoral Candidate, Harvard Business School, Wyss House, Boston, MA, 02163, United States, miketeod@hbs.edu As legal instruments, patents do not follow the linguistical assumptions expected in a natural language. Direct application of standard linguistic tools does not capture the intricacies of the patent corpus. I present a novel corpus model, modified measures of textual similarity, and applications of patent text processing to knowledge flows within and across firms. 20 - Optimization of Scheduling Malleable Tasks on Single-stage Non-identical Machines Qing Wang, Northeastern University, Heping District, 3-11 Wenhua Road, Liaoning, 110819, China, qwang@mail.neu.edu.cn This paper considers a special manufacturing environment of make-to-order quartz processing enterprise. The tasks for scheduling are malleable and each has a due date with penalty for tardiness. The considered production stage is the bottleneck factor of order tardiness in the whole manufacturing process, and there are several equipments or machines in common function at the stage. Each equipment or machine has distinctive processing efficiency and processing costs. The study aims to minimize the total costs in two aspects of the processing costs and the tardiness penalties to find the optimal schedule solution. Bo Han, Texas A&M.University-Commerce, 2600 S.Neal, Commerce, TX, 75428, United States, bo.han@tamuc.edu

196