Reinforcement learning based optimization of process chromatography for continuous processing of biopharmaceuticals

Abstract

Process intensification in the form of continuous processing is presently being adopted by the biopharmaceutical industry as it offers significant advantages over conventional processing. Chromatographic steps form the core separation steps of a typical biopharma process due to their high selectivity and robustness. To this end, this paper proposes a novel approach based on reinforcement learning (RL), wherein a maximization problem is formulated for cation exchange chromatography for separation of charge variants by optimization of the process flowrate. Chromatography analysis and design toolkit have been used for process simulation and the optimum flow rate at which the yield is maximum and purity constraints are satisfied has been estimated based on the reward policy of RL. Results were experimentally validated and indicate that the proposed RL based approach is superior to the conventional trial and error method of optimizing flowrate in terms of both optimality and computational aspects (3X faster).