Plantwide optimization of monoclonal antibody manufacturing platforms via mixed integer nonlinear programming

Abstract Design variables of upstream and downstream process units play a significant role in the performance of monoclonal antibody (mAb) production platforms. Plantwide optimization problems in the form of mixed integer nonlinear programming (MINLP) must consider decision variables, process constraints and a suitable objective function to achieve meaningful solutions. Moreover, deterministic MINLP solvers are required to avoid simulation-based pitfalls. In this study, a deterministic branch-and-bound method employs the BONMIN solver to obtain credible solutions to the non-convex MINLP problem established from a plantwide optimization flowsheet, which simultaneously considers an upstream mAb producing bioreactor and downstream separation design variables. The BONMIN solver has thus been employed here to solve a chromatography sequencing and column sizing problem, where the cost of goods sold metric is optimized, and a minimum annual mAb throughput is achieved. The plantwide MINLP problem that has been formulated integrates upstream bioreactor design, chromatography sequencing and column sizing, simultaneously addressing all for techno-economic optimization. The results obtained are analysed in order to pinpoint the integrated upstream–downstream technologies and operating strategies holding the strongest promise for industrial mAb manufacturing.