Geração de romaneios otimizados para consolidação de cargas na distribuição de mercadorias com a utilização de frota subcontratada

Abstract

This Master’s thesis addresses the freight consolidation integrated to the generation of cargo packing lists, from the operational perspective of an outsourced fleet in road transport with the use of freight FOB (Free on Board). The cargo packing list is a document used in cargo transportation and its purpose is to list the delivery information regarding the invoices of the goods to be dispatched. The problem consists of how to best arrange the loads in the different types of vehicles of the contracted carriers, besides having to respect the collection and redispatch processing capacities, in order to obtain cargo packing lists with the lowest total cost of distribution. These allocation decisions involve a price table, consisting of several price rules, rates and taxes, defined according to the number of deliveries in the cargo packing list and the individual weight of the cargoes to be sent. This causes a piecewise linear behavior in the cost function, requiring its full linearization to be solved as an Integer Linear Programming Problem. This case can be seen as an extension of the classic Bin-Packing Problem, in which each bin to be occupied individually has its utility value manifested as piecewise linear. Motivated by this practical and theoretical context, and because it is also a subject little explored in the literature, the focus of this research is to propose models and solution methods that are effective in supporting decision-making in practice. In addition to solving the proposed models using general-purpose optimization software, we develop constructive heuristics based on these models, as well as on reformulations obtained through the Dantzig-Wolfe decomposition. To verify the adequacy and coherence of these approaches, computational experiments were carried out with real data provided by a manufacturer of school supplies that collaborated in the development of this research. The obtained results revealed a high potential for application of the models for reducing costs in the practice, where reductions of more than 45% were observed in the analyzed scenarios when comparing them with the company's history. The heuristic methods also proved effective for the situation in question, offering quality solutions, yet being 96% faster on average than using exact methods.