Algoritmos de agrupamento capacitado aplicado ao problema de despacho de ordens de serviço

Abstract

With competition increasing, companies must make their operations more lean to avoid unnecessary expenses. For companies that serve customers spread over a city, these services must be made optimally, avoiding waste, both financial and time with unnecessary travels. To avoid such displacement, dispatching of calls for the teams must take into account the location of call in space and time required to perform such call. To promote these benefits, we can look at this problem as a capacitated clustering problem, where we want to associate the calls with the teams so that all calls associated with a team are close to each other, reducing the distance between the calls, and that it is possible they all run within team’s workday. This work is a study of some classical heuristics methods in the literature for the clustering problem and local search apllyed to the service order dispatch problem (SODP), dealing also with some characteristics of the instances as the total capacity surplus and dispersion of care in space. In addition, we developed a small change in one of the existing methods, so that the resulting method could produce better solutions than that obtained with the original method to the vast majority of instances.