A matheuristic decomposition approach for the scheduling of a single-source and multiple destinations pipeline system

Abstract

An improvement on the scheduling of pumping and delivery operations in an installed pipeline network can lead to considerable profits to the using companies, such as oil companies. This paper proposes a decomposition approach that integrates heuristic procedures and mixed integer linear programming (MILP) models, a matheuristic, to solve the long-term scheduling of a pipeline system, which connects a single-source to multiple distribution centers. The approach provides a rigorous inventory management and flow rate control taking into account several operational aspects, such as simultaneous deliveries, and prespecified periods of tank maintenance and pipeline maintenance. To validate the developed approach, two case studies were devised. In case study 1, several instances of an illustrative network were solved and case study 2 addressed three examples of a real-world network: base instance; extended instance with maintenance periods; and model performance tests. Valid solutions that can be operationally implemented were obtained for all executions in a reasonable computational time. Detailed discussions of the obtained solutions are presented and indicate an inventory control in accordance with operational requirements.