Sequencing Batches in a Real-World Pipeline Network Using Constraint Programming

Abstract

Abstract This paper presents a new approach for sequencing of batches in a complex pipeline network by using Constraint Programming (CP) techniques. These batches are composed by oil derivatives to be transferred by the pipeline network. Although many papers found in the literature usually deal with reduced pipeline topologies, the network considered in this work is quite complex. It involves nine areas, including three refineries, one harbor and five distribution centers interconnected by fifteen multiproduct pipelines. Although this paper only focus on sequencing of batches, the final task is to provide a complete transfer scheduling for the pipeline network. This scheduling is accomplished by means of three key elements: assignment, sequencing and timing. The proposed CP model considered for sequencing of batches introduces particular features of the problem that can be used by specialized solvers. This approach aims to consider time delays introduced by the pipeline network during product transfers. Preliminary results show that a suitable ordering of batches can be obtained for improving inventory levels. Besides, results show a better performance when compared to a heuristic approach which can fail for some particular cases.