Novo modelo de programação linear inteira mista para o problema de dimensionamento de frota de aeronaves no transporte aéreo offshore em função do nível de serviço durante crises: pesquisa ação em uma empresa petrolífera

Abstract

Fleet sizing consists of solving the fleet size and composition problems. It can also indicate fleet routing. This work aims to propose a mixed integer mathematical programming model to minimize the total fleet costs, subject to the constraints identified in the literature and with the studied company. The method used in this action research classifies as an applied, quantitative and exploratory research. With the purpose that the fleet sizing occurs according to the company's expectation and at minimum cost, a clear definition about the desired service level is required, in order to explain exactly what the sized fleet must be capable to accomplish. The service level defined and applied in this work through multi-criteria analysis means that the fleet must be sized and the operation properly prepared for the crisis in which: Three aerodromes of the basin remain closed for 1 full day, until the maximum term of 5 days it is required the recovery of flights and operations normalization. To meet this service level, a mixed integer mathematical programming model was written in Lingo software. Although the methodologies are real and this work developed the model with the real problem data, this research used fictitious numbers to demonstrate the mathematical model results, preserving the studied company internal data. This project generated ten different scenarios, covering the different service level alternatives, variations in the macroeconomic scenario, demand, premise and sizing methodology – company’s official versus this research proposal. The results show that the proposed model costs are lower than the methodology used in the studied company, with a reduction from 28% to 33%. The measured results variability was ± 5%. The main contributions are the service level definition for the correct fleet sizing and the weekly periodicity for the effective model mix optimization of the aerial fleet. Similar works in the literature about offshore fleet sizing based on service level for crisis were not found.