Aplicação de programação lógica por restrições para a otimização do sequenciamento de uma rede dutoviária

Abstract

This work uses Constraint Logic Programming (CLP) as a proposal for the activity of sequencing batches in a pipeline network. The developed model is embedded in an optimization framework that aids the scheduling task of transporting, mainly, petroleum derivatives. The pipeline network in study is composed of 14 nodes (or areas), namely: 4 refineries, 2 harbors, 2 final clients, 5 distribution terminals, and a junction of valves and pumps, connected by 30 pipelines. In the above-mentioned network, more than 35 oil products and ethanol can be transported. The process of programming transfer and storage activities involves complex operational constraints, and is characterized as a combinatorial optimization problem of difficult solution. Due to the complexity of the problem, a decomposition strategy is used for modeling purposes. This strategy follows a hierarchical division based on the three key scheduling elements: Allocation of Resources, Sequencing of Activities, and Timing. The model developed in Constraint Logic Programming is applied to the Sequencing of Activities, using a discretized temporal approach. The objective is to obtain the pumping order of batches in origin areas and the sequencing in intermediate nodes, taking into account the management of inventories. The proposed model for the sequencing activities consists of, for instance, restrictions for reversals of flow in pipelines and for insertion of plugs due to incompatible products. The tests were conducted in real-world scenarios for a scheduling horizon of 30 days, with one-hour discretization interval. A comparative study with an MILP (Mixed Integer Linear Programming) model for the same sequencing problem is conducted. The obtained results indicate that the proposed CLP model was competitive to the equivalent MILP model. Thus, CLP was an alternative approach for solving the studied combinatorial problem.