How is linear programming used in oil and chemical refinery blending?

In oil and chemical refinery blending, linear programming is used to determine the most profitable combination of various hydrocarbon streams to produce final products that meet strict quality and market specifications.12

Refineries often manage 8 to 15 different blend stocks, such as butane, naphtha, and reformate, each with unique costs and properties. Simultaneously, the final product must satisfy 10 to 15 quality constraints, including octane ratings, sulfur content, and vapor pressure, alongside local regulatory requirements.2

Linear programming functions by defining an objective function—typically the maximization of operating profit—subject to these linear constraints. By modeling the properties and costs of each component as a system of equations or inequalities, the technique allows refiners to rapidly select the optimal disposition of components from a vast number of feasible alternatives. This mathematical approach ensures that the refinery achieves the highest possible economic return while adhering to complex production and quality limitations.12