This paper is motivated by the need to efficiently solve large-scale production systems with wells comprising both gas-lift and choke control. For near real-time application, the solution must be attainable within a few minutes. However, while previous work (Rashid, 2011) addressed the dual control modes, that is, the use of gas-lift injection to increase the production rate and a choke setting to restrict the rate of flow, it became apparent that with increasing problem dimensionality the time required to reach a solution of the approximate problem grows significantly, rendering the method somewhat impractical. This is a result of the mixed-integer nonlinear formulation adopted, but also due to the burden of managing the fitting and storage of the well performance curves in the approximate model. For these reasons, a bilinear interpolation scheme is presented in the offline step of the iterative procedure to enable a fast solution. The bilinear interpolation scheme is compared with three alternative formulations, including a relaxation of the earlier MINLP model to a piece-wise MILP model. Tests on a real-world production system with over 100 gas-lifted wells demonstrate that the bilinear model, with a fast heuristic for well deactivation, is the most practically viable approach.