[HAS Online Seminar] Reminder: Dr. Hrayer Aprahamian Tomorrow at 1-2pm ET (Friday, 1/27)
8 views
Skip to first unread message
hasseminarseries
Jan 26, 2023, 6:32:32 PM1/26/23
to INFORMS Health Applications Society Online Seminar Series
Dear colleagues,
This is a reminder for our first seminar of the HAS Online Seminar Series this year tomorrow by Dr. Hrayer Aprahamian, from Texas A&M University. The details of the seminar are below. Click here to register for the event!
This is a reminder for our first seminar of the HAS Online Seminar Series this year tomorrow by Dr. Hrayer Aprahamian, from Texas A&M University. The details of the seminar are below. Click here to register for the event!
An Optimization Framework for Customized Targeted Mass Screening of Non-uniform Populations under the Availability of Multiple Schemes and Tests
Friday, January 27, 1-2pm ET (10-11 am PT)
Abstract: In this work, we study the problem of designing optimal targeted mass screening of non-uniform populations. Mass screening is an essential tool that is widely utilized in a variety of settings, e.g., preventing infertility through screening programs for sexually transmitted diseases, ensuring a safe blood supply for transfusion, and mitigating the transmission of infectious diseases. The objective of mass screening is to maximize the overall classification accuracy under limited budget. In this work, we address this problem by proposing a proactive optimization-based framework that factors in population heterogeneity, limited budget, different testing schemes, the availability of multiple assays, and imperfect assays. By analyzing the resulting optimization problem, which is a mixed integer nonlinear programming problem, we establish key structural properties which enable us to develop an efficient solution scheme. To achieve this, we take advantage of a reformulation of the problem as a multi-dimensional fractional knapsack problem and identify an efficient globally convergent threshold-style solution scheme that fully characterizes an optimal solution across the entire budget spectrum. Using real-world data, we conduct a geographic-based nationwide case study on targeted COVID-19 screening in the United States. Our results reveal that the identified screening strategies substantially outperform conventional practices by significantly lowering misclassifications while utilizing the same amount of budget. Moreover, our results provide valuable managerial insights with regards to the distribution of testing schemes, assays, and budget across different geographic regions.
Friday, January 27, 1-2pm ET (10-11 am PT)
Abstract: In this work, we study the problem of designing optimal targeted mass screening of non-uniform populations. Mass screening is an essential tool that is widely utilized in a variety of settings, e.g., preventing infertility through screening programs for sexually transmitted diseases, ensuring a safe blood supply for transfusion, and mitigating the transmission of infectious diseases. The objective of mass screening is to maximize the overall classification accuracy under limited budget. In this work, we address this problem by proposing a proactive optimization-based framework that factors in population heterogeneity, limited budget, different testing schemes, the availability of multiple assays, and imperfect assays. By analyzing the resulting optimization problem, which is a mixed integer nonlinear programming problem, we establish key structural properties which enable us to develop an efficient solution scheme. To achieve this, we take advantage of a reformulation of the problem as a multi-dimensional fractional knapsack problem and identify an efficient globally convergent threshold-style solution scheme that fully characterizes an optimal solution across the entire budget spectrum. Using real-world data, we conduct a geographic-based nationwide case study on targeted COVID-19 screening in the United States. Our results reveal that the identified screening strategies substantially outperform conventional practices by significantly lowering misclassifications while utilizing the same amount of budget. Moreover, our results provide valuable managerial insights with regards to the distribution of testing schemes, assays, and budget across different geographic regions.
Bio: Dr. Hrayer Aprahamian is an Assistant Professor at the Wm. Michael Barnes Department of Industrial and Systems Engineering at Texas A&M University. He received his Ph.D. in Industrial and Systems Engineering from Virginia Tech in 2018. Dr. Aprahamian’s research interests lie at the interplay between combinatorial/discrete and continuous optimization, with particular interest in applications related to healthcare systems and public policy decision-making. His research papers have appeared in journals such as Management Science, INFORMS Journal on Computing, Stochastic Systems, and IISE Transactions. He was awarded the 2022 Pierskalla Best Paper Award, was a finalist for the 2021 JFIG Paper competition, received the IISE Transactions Award in 2020, the Pritsker Doctoral Dissertation Award in 2019, the Paul E. Torgersen Research Excellence Award in 2018, and was runner-up for the 2017 Pierskalla Best Paper Award..
Feel free to share it with anyone who could be interested among your colleagues and student groups.
Looking forward to seeing you there and then!
Best regards,
HAS Online Seminar Organization Team
Sanjay Mehrotra (Northwestern University)
Looking forward to seeing you there and then!
Best regards,
HAS Online Seminar Organization Team
Sanjay Mehrotra (Northwestern University)
Qiushi Chen (Penn State)
Lauren Steimle (Georgia Tech)
Reply all
Reply to author
Forward
0 new messages