The vehicle routing problem (VRP) arise from the need to promote efficiency in the distribution of products or services between customers, suppliers, partners, and other. The VRP variant of interest in this project addresses joint pickup and delivery operations with customer time windows (PDPTW). Mathematical models and solution methods proposed in the literature assume that travel times and demand are previously known and fixed during the routes execution, characteristics which are not practical. Therefore, aiming to approach the problem in a more realistic way, the objective of this project is to propose models and algorithms for the PDPTW in which is considered travel times and demand uncertainties. For this, it is intended to apply the Robust Optimization (OR), in which allows to include in the problem sets of uncertainties which model the variation of uncertain parameters. Include uncertainties in the PDPTW through OR is not a trivial extension because the demands can be paired in customers and the vehicle load is not necessarily monotonic during the route as in other vehicle routing probelms, which make it difficult to handle the uncertainties by current approaches. Therefore, it is intended to contribute with innovative models and solution methods for the PDPTW, which could also be applied or extended to other variants with pickup and delivery.
News published in Agência FAPESP Newsletter about the scholarship: