Optimal Control Strategy of Multi-Stage Pharmaceutical

Abstract

Abstract The fermentation production process is the basis for ensuring product quality in pharmaceutical production. There are many variable parameters such as temperature, number of revolutions and PH value in the fermentation production process, therefore, it is difficult to describe the complex problems through precise mathematical models. Since the process has to go through four different steps, each of which is a nonlinear time-varying dynamic system, the four steps constitute a large dynamic system, and the global modeling accuracy cannot be guaranteed; and the optimization problem of the nonlinear time-varying dynamic system is also a problem. In order to solve the above problems, the paper proposes to use the BP (Back Propagation）neural network modeling method to screen the input variables of the system with large effect on the basis of the actual production data of a pharmaceutical factory. Then, based on the process parameters of the fermentation and the output of the product quality index, a multi-stage mapping model based on LSSVM (Least Squares Support Vector Machine) was established. Finally, the PSO (Particle Swarm Optimization) algorithm is used to optimize the parameters of the model.The simulation results show that the model obtained by parameter optimization can achieve better prediction results.