About

Serhan Ziya is a Professor at the University of North Carolina at Chapel Hill, working within the Department of Statistics & Operations Research. He holds a B.S. in Industrial Engineering from Bogazici University, obtained in 1997, and both an M.S. and Ph.D. in Industrial and Systems Engineering from Georgia Institute of Technology, earned in 1999 and 2003 respectively. His research primarily focuses on stochastic modeling, healthcare operations, service operations, queueing design and control, and revenue management. He works on problems that are directly motivated by specific applications such as healthcare, utilizing techniques from stochastic modeling and operations research. He emphasizes the importance of developing practically sound and useful solutions to complex problems through creative approaches that integrate various methods. To this end, he often collaborates with statistician colleagues and domain experts to enhance the applicability and effectiveness of his research.

Research topics

• Computer Science
• Medicine
• Mathematics
• Artificial Intelligence
• Emergency medicine
• Engineering
• Real-time computing
• Economics
• Mathematical optimization
• Computer network
• Business
• Operations research
• Operations management
• Internal medicine
• Intensive care medicine
• Statistics
• Medical emergency
• Nursing

Selected publications

• Frontiers in Operations: Show-Up Profiles for Scheduled Services: Estimation and Applications

  Manufacturing & Service Operations Management · 2026-02-18

  articleSenior author

  Problem definition: Motivated by passenger arrivals at the security checkpoint of the Raleigh-Durham International Airport, we develop methods to study arrivals to a system in which they are tied to scheduled events, such as flights. A key concept for modeling arrivals in such systems is the “show-up profile,” a probability distribution describing how far in advance passengers arrive for their flights. These profiles can be combined based on a known flight schedule to yield an aggregate passenger arrival forecast. Existing industry practice and academic work estimate show-up profiles using surveys or other data that are typically not available to U.S. airports. This motivates our study of an easy to implement and dynamic method for estimating show-up profiles. Methodology/results: We introduce an innovative solution for estimating show-up profiles using infrared-beam people-counting sensors and a structural estimation approach that does not require a mapping of passengers to flights. A direct maximum likelihood approach is intractable, but we propose a tractable approximation and prove that it yields consistent estimates of the underlying show-up profile parameters. Our approach produces forecasting results comparable to pure machine learning methods, yields significantly improved adaptive forecasts when combined with machine learning methods, and reveals empirical insights about passenger behavior variations across different times of day and flight destinations. Managerial implications: Our work presents a novel application of Internet of Things technology to service operations with incomplete data and demonstrates the value of integrating known operational structure with black box forecasting approaches. Show-up profiles are used at airports for decision making, for example, for crowd management, and our methodology has the potential to drive significant improvements in airport operations. The methods we develop can be readily applied at U.S. airports and other transportation hubs, and they can be adapted to other event-driven service environments such as theaters, healthcare facilities, and museums. History: This paper has been accepted in the Manufacturing & Service Operations Management Frontiers in Operations Initiative. Funding: This work was supported by the 2021 Triangle Impact Challenge. Supplemental Material: The online appendices are available at https://doi.org/10.1287/msom.2024.1575 .

  PublisherDOI

• Dynamic Distribution of Patients to Medical Facilities in the Aftermath of a Disaster

  UNC Libraries · 2026-04-08

  articleOpen access

  In the aftermath of a disaster, emergency responders must transport a large number of patients to medical facilities, using limited transportation resources (such as ambulances). Decisions about where to send the patients are typically made in an ad hoc manner by responders on the scene. Using a Markov decision process formulation, we develop two heuristic policies that use limited information such as mean travel times and congestion levels to determine (a) how to allocate ambulances to patient locations and (b) which medical facility should be the destination for those ambulances. In a simulation study, we incorporate patient survival rates and service times for different types of traumatic injuries, and show that the proposed heuristics can provide substantial improvement in the expected number of survivors compared to the common practice of transporting to the nearest facility, even when the decision maker has only limited up-to-date information about the system state. In particular, a myopic approach that considers only what is best for the next patient to be transported increases the expected number of survivors in almost all scenarios considered. Using a more sophisticated one-step policy improvement approach provides further improvement when the event involves patients who do not deteriorate rapidly, especially when the transportation is not the bottleneck and the casualties are spread over many locations. We demonstrate the effectiveness of the proposed heuristics on a case study of a hypothetical earthquake, where casualty data is generated using computer software developed by the U.S. government. The e-companion is available at https://doi.org/10.1287/opre.2017.1695 .

  PublisherDOI

• Comparison of emergency department crowding scores: a discrete-event simulation approach

  UNC Libraries · 2026-04-15

  articleOpen access1st authorCorresponding
  PublisherDOI

• When to Triage in Service Systems with Hidden Customer Class Identities?

  UNC Libraries · 2026-04-14

  articleOpen access1st authorCorresponding

  In service systems with heterogeneous customers, prioritization with respect to the relative importance of customers is known to improve certain performance measures. However, in many applications, information necessary to determine the importance level of a customer may not be available immediately but can be revealed only through some preliminary investigation, which is sometimes called triage. This triage process is typically error‐prone and may take substantial amount of time, and hence, it is not always clear if and when it should be implemented for purposes of priority assignment. To provide insights into this question, we study a stylized queueing model with a single server and two types of customers with hidden type identities, which differ in their rates of service and waiting costs. By means of a Markov decision formulation, we first show that the optimal dynamic policy on triage is characterized by a switching curve. The comparison of two state‐independent policies (no‐triage and triage‐all) shows that the information from triage is more beneficial when the traffic intensity is neither too low nor too high. Our numerical results show that the system manager should consider implementing a state‐dependent triage policy when the probability of classifying a customer into the important class and the mean triage time are of moderate size, when the difference between the importance levels of the two classes of customers is large, and/or when the traffic intensity is high.

  PublisherDOI

• Patient Triage and Prioritization Under Austere Conditions

  UNC Libraries · 2026-04-08

  articleOpen accessSenior author

  In war zones and economically deprived regions, because of extreme resource restrictions, a single provider may be the sole person in charge of providing emergency care to a group of patients. An important question the provider faces under such circumstances is whether or not to perform triage and how to prioritize the patients. By choosing to triage a particular patient, the provider can determine the health condition and thus the urgency of the patient, but that will come at the expense of delaying the actual service (stabilization or initial treatment) for that patient as well as all the other patients. Motivated by this problem, which also arises in other service contexts, we consider a service system where finitely many patients, all available at time zero, belong to one of the two possible triage classes, where each class is characterized by its waiting cost and expected service time. Patients’ class identities are initially unknown, but the service provider has the option to spend time on triage to determine the class of a patient. Our objective is to identify policies that balance the time spent on triage with the time spent on service by minimizing the total expected cost. We provide a complete characterization of the optimal dynamic policy and show that the optimal dynamic policy that specifies when to perform triage is determined by a switching curve, and we provide a mathematical expression for this curve. One insight that comes out of this characterization is that the server should start with performing triage when there are sufficiently many patients and never perform triage when there are few patients. Finally, we carry out a numerical study in which we demonstrate how one can use our mathematical results to develop policies that can be used in mass-casualty triage and prioritization, and we find that there are substantial benefits to using one of these policies instead of the simpler benchmarks. The online supplement is available at https://doi.org/10.1287/mnsc.2017.2855 . This paper was accepted by Assaf Zeevi, stochastic models and simulation.

  PublisherDOI

• Patient sex, racial and ethnic disparities in emergency department triage: A multi-site retrospective study

  UNC Libraries · 2025-04-03

  articleOpen access
  PublisherDOI

• Test Allocation and Pool Composition in Heterogenous Populations Under Strict Capacity Constraints

  Manufacturing & Service Operations Management · 2025-06-27

  articleSenior author

  Problem definition: Motivated by the persistent lack of testing capacity in the first year of the COVID-19 pandemic, we study the question “who should be tested?” when there are general costs and rewards, testing capacity is strictly limited, tests have errors, and patients differ in their prior probability of being infected. We specifically study how the answer to that question changes when pooled testing, a method of grouping samples to conserve tests, is an option. Methodology/results: We use a two-stage stochastic optimization model with recourse, incorporating costs and rewards for different test outcomes, under a conservative capacity constraint that reflects severe shortages of tests or high uncertainty about future test availability. This setting reflects the situation decision makers faced at the beginning of the COVID-19 pandemic in March 2020. Although health officials might intuitively prioritize testing patients who are highly likely to be infected, we find that it may be better to focus on patients who are less likely to be infected, particularly when the test has low sensitivity (i.e., the false-negative rate is substantial). Moreover, it may be optimal to test two groups of individuals: those who are very unlikely to be infected (in pools) and those who are very likely to be infected (individually). Managerial implications: We develop a heuristic policy supported by the analysis, which indicates when pooling should be used and which type of samples should be tested. In some settings, the decision may be characterized simply by understanding the costs and rewards involved. In more complex testing settings, the characteristics of the test and the size of the pool affect the desirability of pooling: Lower specificity, higher sensitivity, and larger pool sizes all result in testing environments that are more favorable to pooling. Managers and policymakers should understand how characteristics of the test and the setting impact whether it is optimal to test patients who are deemed likely to test positive or those who are likely to test negative. Incorporating pooling as a test strategy may change which patients should be prioritized for a test. Our results can inform both public health policy and healthcare operations management in settings where testing capacity is strictly limited. Funding: This work was supported by the Division of Civil, Mechanical and Manufacturing Innovation [Grant 1634822]. S. Ziya was supported by the National Science Foundation [Award CMMI1635574]. Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2021.0149 .

  PublisherDOI

• Show-Up Profiles for Scheduled Services: Estimation and Applications

  SSRN Electronic Journal · 2025-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Dynamic Resource Allocation in Urban Search and Rescue Operations

  SSRN Electronic Journal · 2024-01-01

  preprintOpen accessSenior author
  PublisherDOI

• Identifying Patient Subpopulations with Significant Race-Sex Differences in Emergency Department Disposition Decisions

  Health Services Insights · 2024-01-01

  articleOpen accessSenior author

  Background/objectives: The race-sex differences in emergency department (ED) disposition decisions have been reported widely. Our objective is to identify demographic and clinical subgroups for which this difference is most pronounced, which will facilitate future targeted research on potential disparities and interventions. Methods: We performed a retrospective analysis of 93 987 White and African-American adults assigned an Emergency Severity Index of 3 at 3 large EDs from January 2019 to February 2020. Using random forests, we identified the Elixhauser comorbidity score, age, and insurance status as important variables to divide data into subpopulations. Logistic regression models were then fitted to test race-sex differences within each subpopulation while controlling for other patient characteristics and ED conditions. Results: In each subpopulation, African-American women were less likely to be admitted than White men with odds ratios as low as 0.304 (95% confidence interval (CI): [0.229, 0.404]). African-American men had smaller admission odds compared to White men in subpopulations of 41+ years of age or with very low/high Elixhauser scores, odds ratios being as low as 0.652 (CI: [0.590, 0.747]). White women were less likely to be admitted than White men in subpopulations of 18 to 40 or 41 to 64 years of age, with low Elixhauser scores, or with Self-Pay or Medicaid insurance status with odds ratios as low as 0.574 (CI: [0.421, 0.784]). Conclusions: While differences in likelihood of admission were lessened by younger age for African-American men, and by older age, higher Elixhauser score, and Medicare or Commercial insurance for White women, they persisted in all subgroups for African-American women. In general, patients of age 64 years or younger, with low comorbidity scores, or with Medicaid or no insurance appeared most prone to potential disparities in admissions.

  PublisherDOI

Recent grants

• Collaborative Research: In Search of the Greatest Good with Imperfect Triage in the Aftermath of Mass-Casualty Events

  NSF · $390k · 2012–2016

• Collaborative Research: Patient Triage in the Aftermath of a Mass Casualty Event - A Dynamic Programming Approach

  NSF · $156k · 2006–2010

Frequent coauthors

• Nilay Tanık Argon

  University of North Carolina at Chapel Hill

  28 shared

• Benjamin Linthicum

  University of North Carolina at Chapel Hill

  11 shared

• Yufeng Liu

  9 shared

• Christopher S. Evans

  East Carolina University

  8 shared

• Qian Cheng

  Chinese Academy of Agricultural Sciences

  7 shared

• Robert D. Foley

  6 shared

• Mehul D. Patel

  University of North Carolina at Chapel Hill

  6 shared

• Peter Lin

  University of North Carolina at Chapel Hill

  6 shared