Optimisation Studies of Fluorination Reactions Based on Bayesian Optimisation Algorithm

The fluorination reaction has a wide range of applications in the synthesis of drugs and pesticides. Bayesian optimisation can construct the mathematical connection between the key experimental parameters and the product yield, and rapidly optimise the sieve of process parameters and conditions. The establishment of a comprehensive objective function allows for the simultaneous consideration of multiple optimisation objectives and constraints of the experimental parameters, and the final determination of the influence ranking of the seven input variables on the predicted output of the model. The optimal combination of experimental parameters was identified, resulting in an error (ARD%) of 1.38 % between the experimental value (0.6004) and the predicted value (0.6087). This optimisation model demonstrates a commendable predictive capability, surpassing the efficacy of conventional optimisation methods.