Mixed-integer linear optimization for optimal lift-gas allocation with well-separator routing

The lift-gas allocation problem with well-separator routing constraints is a mixed-integer nonlinear program of considerable complexity. To this end, a mixed-integer linear formulation (compact) is obtained by piecewise-linearizing the nonlinear curves, using binary variables to express the linearization and routing decisions. A new formulation (integrated) combining the decisions on linearization and routing is developed by using a single binary variable. The structures of both formulations are explored to generate lifted cover cuts. Numerical tests show that the solution of the integrated formulation using cutting-plane generation is faster in spite of having more variables than the compact formulation.

► Models for lift-gas allocation in oil wells with well-separator routing.
► MILP models obtained by piecewise linearizing the nonlinear well production curves.
► Cutting planes based on knapsack covers for the MILP models.
► A computational analysis comparing the models and impact of cutting planes.