Gestão de pessoas em navios da esquadra: uma proposta de métodos quantitativos para a distribuição de vagas em cursos para os militares que servem embarcados, utilizando programação linear inteira, opinião de especialista, simulações estocásticas e medidas de risco

Abstract

For a Navy to be ready, it is necessary not only to have technologically advanced assets, but also qualified personnel to operate them. It has been observed that the allocation of vacancies in operational courses common to ship crews is not supported by quantitative methods. This work proposes methods to support this annual allocation for service members embarked on Fleet ships. Monthly series from 2020–2023 are used for a sample of 18 courses common to three ships from the same Squadron, integrating forecasting of count time series with risk-oriented stochastic optimization and expert opinion. The series are modeled as count processes of the random-walk type with Bernoulli–Poisson terms, generating 12-month scenarios validated by fan charts and expert judgment. Five integer linear programming models are formulated: a deterministic model; a model with a buffer defined by an expert; and three stochastic models based on simulations and on Value at Risk (VaR) and Conditional Value at Risk (CVaR). All of them yielded a global cost lower than that of the actual allocation. The deterministic model achieved the best trade-off between cost and coverage; the buffered model increased the cost but added robustness; the VaR- and CVaR-based models were more cost-efficient, but with lower initial coverage. No course fell below the minimum index in any of the suggested allocations. It is concluded that the methods are operationalizable and reproducible, allowing the calibration of risk and buffer to better meet on-board training needs and contribute to the readiness of the ships.