Mulesa O. Information technology of forecasting and client-oriented personnel optimization of health care facilities.

The dissertation provides an overview of current approaches to personnel optimization of health care facilities in Ukraine. The research findings show that there is no universally accepted integrated approach to the solution of this problem. The dissertation provides rationale for the development of intelligent knowledge-based systems to support decision making when implementing programmed human resource (HR) policy in health care facilities. Programmed HR policy is defined as a set of steps that are taken by the management of health care facilities in order to improve the quality and timeliness of the delivery of medical services in the future. The dissertation establishes the framework for the conception of programmed HR policy in health care facilities. The research has found that the implementation of the effective programmed HR policy requires accurate and timely forecasting data on the demand for health services. The findings of this research show that the prerequisite to the effectiveness of HR policy is accurate and timely forecasting of the demand for medical services in future time periods. The dissertation contains the sequence of the data analysis steps that has been developed in order to evaluate the future demand for medical services and provides rationale for it. The steps include clustering, identification and forecasting. The method of evolutionary clustering and the adaptive method of fuzzy c-means have been developed to solve the problem relating to clustering people who are potential consumers of medical services. The application of these methods makes it possible to increase the accuracy of clustering and to take into account the fuzzy non-numerical characteristics of individuals when dividing them into groups. This research suggests that the problem of forecasting the demand for medical services on the basis of time series be solved with the methods of synthesis of forecast schemes developed in this study on the basis of basic forecast models, which improves the volatility of the obtained forecast values. The results indicate that the method of fuzzy classification that was developed based on Wald's sequential analysis enables to predict the demand for medical services based on individual characteristics of individuals and can become a tool for healthcare schedule management. This research suggests that the methods of structural and parametric identification of unknown tabular relationships that make it possible to obtain adequate models be used to determine the relationships between the number of groups of potential consumers of medical services and the demand for such services. The obtained forecast data on the demand for medical services in the future serve as a basis for making relevant management decisions on the programmed HP policy in health care facilities. In order to settle these decision-making issues fuzzy voting methods and methods of numerical evaluation of objects have been developed. They make it possible to take into account the subjective nature of expert opinions. The established criteria for selecting optimal alternatives involved in multistage decision making process provide for taking into account the effects of decision making at the subsequent stages of their implementation.