Scheduling medical appointments plays a fundamental role in managing patient flow and ensuring high-quality care. However, no-shows can significantly disrupt this process and affect patient care. To address this challenge, healthcare facilities can adopt different strategies, including overbooking in medical consultations. While this reduces the risk of unused slots, it can generate associated costs and affect the perception of service quality. In this article, we propose an integer linear optimization model that maximizes the expected utility of a medical center, considering the risk of no-shows and overbooking. For this purpose, machine learning is used to estimate the propensity of each patient to attend their medical appointment, using real data from three medical specialties of a hospital. The results of the application demonstrate the model’s ability to assign appointments and perform overbooking efficiently and in an organized manner, implying an improvement in the utility of a medical center and a positive impact on the perception of the quality of care.

The scheduling of medical appointments is an activity of great importance in a hospital since different human and material capitals must be used efficiently. One of the problems of this work is the non-attendance of a patient, which decreases the efficiency of the use of resources. Several studies have proposed considering "absenteeism" for scheduling medical appointments to address this. However, predicting it is a complex task. This research proposes the prediction of absenteeism to medical appointments for three medical areas of the Hospital Clínico Regional Dr. Guillermo Grant Benavente in the city of Concepción, Chile. For this purpose, five Machine Learning algorithms are trained and evaluated. The best-trained model managed to be a predictive tool of a patient's absenteeism level for his next appointment and to characterize those patients with higher levels of no-show.

Intelligent management of medical appointments can significantly enhance patient care and reduce health risks. By leveraging disease and attendance propensity, our system aims to minimize the likelihood of patients developing serious conditions due to missed appointments by strategically scheduling those at higher risk. For its development, the methodology begins with the formulation of an optimization model, using assumed values for disease and attendance propensity to establish an initial, efficient scheduling of appointments. Subsequently, machine learning algorithms are applied to patient historical data to obtain more precise and realistic estimates of these propensities, which are then integrated into the model to adjust appointment allocation according to each patient’s individual risk. The results demonstrate that the Intelligent Medical Appointment Management Model significantly outperforms random scheduling, which simulates current real-world practices without the use of an intelligent patient assignment system. Patients scheduled by the intelligent model show higher mean propensities to attend appointments and to develop diseases, ensuring that medical resources are allocated efficiently to those in greatest need. Statistical validation confirms the model’s effectiveness, showing significant differences in scheduling outcomes between the intelligent and random models. This approach highlights the potential to reduce health risks among a group of patients by utilizing both their medical histories and synthetic data for more accurate predictions and effective scheduling.

This study investigates the worst-case performance guarantee of locally optimal solutions to minimize the total weighted completion time on uniformly related parallel machines. The investigated neighborhood structure is Jump, also called insertion or move. This research focused on establishing the local optimality condition expressed as an inequality and mapping that maps a schedule into an inner product space so that the norm of the mapping is closely related to the total weighted completion time of the schedule. We determine two new upper bounds for the performance guarantee, which take the form of an expression based on parameters that describe the family of instances: the speed of the fastest machine, the speed of the slowest machine, and the number of machines. These new bounds outperform the parametric upper bound previously established in the existing literature and enable a better understanding of the performance of the solutions obtained for the Jump neighborhood in this scheduling problem, according to parameters that describe the family of instances.

This article introduces, models, and solves the time-dependent in-building delivery problem in last-mile logistics. It determines efficient travel sequences for a worker (e.g., delivery person, deliveryman, mailman, agent) who delivers goods or provides services directly to customers located within a building using its elevation system. We study, in detail, all the steps involved in a travel sequence inside a building: horizontal trips, unloading products to the customers, waiting for elevators, and vertical trips within elevators. The sequences and their total times vary depending on the building type, the elevation system, the moment of the day, and the arrival time because of the daily building traffic intensity variations. A mixed-integer linear programming model and a genetic algorithm-based metaheuristic are proposed to solve a set of instances in two office buildings. The results show that it is very important to determine the best time to visit a building because of its time dependency. The variation in delivery time between off-peak hours versus peak hours is between 15% and 30% for the set of solved instances. Moreover, the order of customer visits differs drastically depending on the arrival time to the building.