Using MOGA to order batches in a real world pipeline network

Abstract

The scheduling of activities to distribute oil derivate products through a pipe network is a complex combinatorial problem that presents a hard computational solution. This problem could be decomposed on three sub problems according to the key elements of scheduling: assignment of resources, sequencing of activities, and determination of resource timing utilization by these activities. This work develops a model to the sequencing sub-problem. The main objective is to develop a multi-objective genetic algorithm to order oil derivate products batches input into the network. From the operational practice, the batches sequencing has great influence on the final scheduling result. The MOGA model provides a set of solutions that means different options of pipeline operations, in a small computational time. This work contributes to the development of a tool to aid the specialist to solve the batch sequencing problem, which reflects in a more efficient use of the pipeline network. © 2010 Springer-Verlag.