04:30 PM - 05:00 PM
A SARIMA Modeling Tool Applied for Forecasting Platelet Bags
This paper describes a systematic approach for automatic identification of parametric models of time-series with forecasting purposes. A general Seasonal Autoregressive Integrated Moving Average (SARIMA) model based on Box & Jenkins procedure is considered in this study. This automatic approach helps the user to identify the degree of non-stationary and seasonal differences in order to fit a complex original series into a stationary series and, as well as to identify orders and parameters of this general SARIMA model. In this study, mathematical models identified from this automatic approach are used to forecast quantities of platelets bags that are need to be distributed from a blood center to the blood bank of hospitals. An example illustrates such an application.
05:00 PM - 05:30 PM
A Solution Algorithm to Surgery Scheduling and Surgeons Assignment in a Public Hospital
The purpose of this study was to solve the weekly surgery scheduling and the surgeon assignment problem in a public hospital that treats indigent patients. The study was divided into two parts: first, surgeries were scheduled with support from a multi-knapsack mathematical model; then, surgeons were assigned, using a search method based on chronological backtracking heuristic. The computer implementation used ILOG libraries in C++ language, obtaining surgery scheduling in minimal time. When we compare the results our algorithm to the current manual procedure, our algorithm has 17.53% more surgeries in some cases. Additionally, the new tool resulted in a homogeneous assignment of surgery among surgeons with the same specialty.
05:30 PM - 06:00 PM
Patient Scheduling for Multiple Services of Cancer Supportive Care
Year by year, the number of cancer patient evolves worldwide. Trends are confirmed in terms of decreasing cancer mortality and increasing gains in life expectancy after cancer. Therefore we are confronted now with the new challenge of the construction of coordinated, customized care process to each patient during and after cancer treatment. The main objective of this research work is to identify better admission criteria that drive optimized efficiency in cancer supportive care services and bring more convenience to the patients. A new decision support tool has been proposed through linear and nonlinear programming to optimize patient recruitment and services coordination for cancer supportive care with minimal impact on the existing organization. Furthermore, an event-driven mechanism is developed to drive the mathematical models, in order to improve the agility of this tool. Numerical examples show that the application of our mathematical model results in significant improvement in admitted list, compared with current scheduling process, especially for a complex case.