| Summary: | In the upstream logistics chain of the oil and gas industry, the supply vessel routing problem in offshore platform groups possesses various variants. The problem variant treated in this work is concerned with the routing of supply vessels whereby two types of service requests are allowed: planned service requests, which are placed prior to supply vessel departure from base port, and random service requests, which are placed during supply vessel tour. In view of the complexity of the stochastic dynamic vehicle routing problem (SDVRP) to which the problem variant studied in this paper belongs, the contribution of this paper is in the mathematical modeling approach and the proposed two-step mixed integer linear programming (MILP) reoptimization heuristic. Assessment of the computational heuristics performance is carried out using a real world case study based on industry practice in the South Atlantic Ocean. In this evaluation for the values of degree of dynamism that is encountered in industrial practice, the competitive ratio of all considered instances is not far from the unit value, a fact that proves the quality of the heuristic employed in this work. Na cadeia logística upstream da indústria de óleo e gás, o problema de roteamento de embarcações supridoras para grupos de plataformas possui diversas variantes. A variante tratada neste trabalho diz respeito ao roteamento destas embarcações, no qual dois tipos de solicitações de serviço são permitidos: solicitações de serviço planejadas, as quais ocorrem antes do zarpe da embarcação de suprimento da base portuária, e solicitações de serviço aleatórias, as quais ocorrem durante a viagem da embarcação de suprimento através das plataformas clientes. Diante da complexidade do problema estocástico de roteamento dinâmico de veículos (SDVRP - Stochastic Dynamic Vehicle Routing Problem) ao qual pertence a variante estudada, a contribuição deste trabalho está na modelagem matemática e na heurística de reotimização de programação linear inteira mista em duas etapas. O ...
|
|---|