Uma análise usando teoria de filas do problema de carregamento de pedidos no porto para abastecimento de unidades marítimas de exploração e produção de petróleo

Abstract

This thesis deals with a supply system comprised of a maritime installation and a fleet of vessels that work together to deliver supply requests to offshore installations. The problem was formulated by a GI/M/S model that considers the port environment, that is, the berths are the servers, and an M/M/S model that considers the maritime environment, that is, vessels are the servers. To address the GI/M/S model, an embedded Markov chain approach was proposed whereby the states space was divided into two subsets, one representing the fully busy system and the other accounting for the situations when the system is partially idle. That representation enabled to find an analytical solution for the fully busy system, which can be combined with simple absorption analysis in the remaining states to derive the exact solution to the steady state probability equations. This contrasts with the classical approach of truncating the state space to find the approximate steady state probabilities. The M/M/S approach at the second installation enabled to address the fleet sizing problem and find solutions based on the occupancy rates of the vessels and the waiting time for impending requests.