Seri halde iki kuyruktan oluşan bir hastanenin hasta kabul kontrolü

The prolongation of human life in parallel with the developments in medicine increases the number of patients who need intensive care. Costs of critical care medicine are 13.4% of hospital costs and 0.66% of the gross domestic product in United States. Despite these expenditures, 90% of intensive care units do not have enough capacity. Under these circumstances, most intensive care units face high occupancy rates. These high occupancy rates might cause people in need of intensive care to be rejected and increase the risk of mortality. An efficient admission control policy may enhance the service processes significantly. In this research, two Markov decision process models which represent intensive care units and related wards in tandem are provided. Both models consist of two types of patients who are intensive care unit and ward patients. In model-1, recovery process of patients who are blocked in the ICU due to lack of available beds in the ward continues, while in model-2 this process is assumed to stop till there is an available bed in the ward. Using these models, the effects of the interaction between wards and intensive care units on optimal admission policies can be understood. Preferred class is a group of patients who are always admitted when there is an available bed. Certain special conditions must be met for the existence of preferred classes. I present sufficient conditions for each patient type to be preferred in both models. At the end, numerical analyses are presented to demonstrate the characteristics of the models. A variation of model-1 is also included in the numerical analysis to observe the effect of an additional type of patients, called post-operational patients, on the system.