Please use this identifier to cite or link to this item:
Title: Diseño de un método meta-heurístico para resolver el problema de asignación de turnos de enfermería ( NSP) con soft-constraints
Authors: Giedelmann Lasprilla, Nicolás
Escuela Colombiana de Ingeniería Julio Garavito
Keywords: Optimización combinatoria
Procesamientos de datos
Issue Date: 2017
Abstract: Three mathematical models based on integer programming and a heuristic method based on iterated local search are proposed to solve the nurse scheduling problem for a hospital taking into account the burnout suffered by nurses caused by continuous working hours and variations in the assigned shifts during consecutive working days. A homogeneous set of nurses must be scheduled on a set of 8 hours working shifts: Morning, Afternoon and Night, or a 24 h “day-off”. The planning is performed for a finite number of days and it prevents the burnout of nurses by trying to schedule constant shifts and thus, the staff is able to regulate the sleep cycles. In addition, nurses are limited to work up to three days in a roll. The objective function is to minimize violations of a preset set of institutional policies of hospitals that are desirable, but not mandatory. These include scheduling a minimum percentage of nurses per shift per period to ensure a good service level. the performance of the proposed approaches is compared on a set of 500 randomly generated instances with up to 500 nurses and 7 days. Results show that one of the mathematical models has better performance than the others and the heuristic method is competitive since it finds solutions with an optimality gap of 7% computed in 0.3s
Appears in Collections:ED - Trabajos de Grado Maestría en Ingeniería Industrial

Files in This Item:
File Description SizeFormat 
TG-08_Nicolas Giedelmann Lasprilla.pdfTesis para optar por el título de mágister en ingeniería Industrial1.2 MBAdobe PDFView/Open
escanear0002.pdfAutorización firmada1.53 MBAdobe PDFView/Open

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.