Afetler sonucunda gözlemlenen sağlık yönetimi sorunlarına yönelik rassal optimizasyon uygulamaları

Disasters and healthcare management are interconnected, often manifesting in a surge of demand for healthcare services, thereby posing difficulties in managing the healthcare operations. This thesis addresses two specific healthcare management problems that arise during disasters. For both problems stochastic optimization approach is adopted to allow incorporation of problem-specific uncertainties into the decision making process. The first part of the thesis centers on the development of a long-term applicable pandemic vaccination calendar. The addressed problem is prompted by the challenges observed during the COVID-19 pandemic, where public health authorities grappled with the development and maintenance of vaccination schedules. The frequent changes in vaccination programs not only drew criticism toward governments but also resulted in significant delays in inoculating the population and restoring normalcy. Our aim is to provide decision makers a mathematical modeling based framework on how can long-term applicable calendars can be developed. Our mathematical approach accounts for pandemic-specific features such as priority groups, different dosing schemes of multiple vaccine types, capacity limitations, demand and uncertainties in vaccine supplies and arrival times-that emerged as critical concerns during the pandemic. Additionally, we extend our study for a reactive approach, allowing updates to calendar decisions in response to mandated changes such as removal of some vaccines or changes in dosing schemes, as observed during the COVID-19 pandemic. The second part of the thesis focuses on the design of a centralized and alliancebased coordination mechanism for renal relief efforts. Previous disasters have consistently highlighted gaps in disaster preparedness and response levels for patients with chronic kidney disease, whose reliance on regular treatment makes interruptions potentially fatal. Following disasters that render dialysis clinics inoperable, swiftly transferring these patients to available ones become a critical concern. Currently, public health authorities collaborate with clinics to facilitate these transfers from shelters, but face challenges in identifying available clinics and matching patients accordingly, leading to ad-hoc transfers that complicate the process. To address this, we propose a mathematical modeling approach to assist a centralized decision maker in establishing alliance relationships among clinics and assigning backup clinics for patients before disasters occur. Our approach accounts for uncertainties in post-disaster clinic capacities, ensuring the relevance of pre-disaster decisions in the aftermath of disasters. Ultimately, our study aims to equip public health authorities with the necessary skills for effective disaster preparedness and response phases in renal relief efforts.