Heurísticas para transporte de pacientes dentro de hospitais

Abstract

In this research, heuristic methods are developed to determine vehicle routes transporting patients inside a hospital. The problem concerns the picking and delivering of patients among the hospital's units. The transportation sector of the hospital receives the requests, which arrive during the day in real time. These contain the patients' information, including the procedures necessary for the care. When a request is available, a decision should be taken, observing the vehicle's disponibility and the existing routes since the patients should be picked up and delivered within a time window. The objective is to minimize the total weighted tardiness in serving all requests. The tardiness occurs when a patient is delivered after the end of her time window. The tardiness is weighted by the patient's urgency. As a way to obtain an effective solution to the problem, a re-optimization heuristic is proposed that uses different rules for decision-making. The rules are based on how patients are chosen for picking up and delivery to attain the problem's objective. The rules take into consideration available information about patients and/or the hospital. Furthermore, a simulated annealing metaheuristic is developed to be used by the re-optimization heuristic. The developed methods are compared concerning the problem's objective, the number of vehicles, and the vehicle's capacity, considering randomly generated instances and also realistic instances obtained from other papers in the literature. The results indicate that among the rules used, those based on the distance (time) for the vehicle to move from one unit to another are more favorable to reducing the total weighted tardiness and, thus, obtaining quality solutions. On the other hand, the best results are obtained when simulated annealing is used for decision-making. Concerning the number of vehicles and their capacity, results indicate that the greater the number of vehicles and their capacity, the smaller the total weighted tardiness.