Constraint Logic Programming Applied to Sequencing Tasks in a Pipeline Network

Abstract

This article applies Constraint Logic Programming (CLP) for the sequencing of operational tasks, namely, the pumping of batches, in a real-world pipeline network that transports oil derivatives and ethanol. This network is located in southeast Brazil and involves refineries, harbors, and final clients, interconnected by 30 pipelines of different sizes, many of them with various kilometers long. The solution of pipeline-scheduling problems poses significant computational challenges. A decomposition framework is used for modeling and solving purposes. The proposed CLP solution approach is used for the combinatorial task of sequencing the operational batches, taking into account a series of operational constraints that are detailed modeled. In particular, the inclusion of plug products is avoided and the minimization of reversion on flow direction of pipelines is taken into account. Tests were conducted based on 30-day scheduling scenarios and operational answers were obtained.