Modeling and optimized controlling of fermentation process based on QPSO and LSSVM

In accordance with optimization control and modeling of polymyxin fermentation process, least square support vector machine(LSSVM) model was established, and a method was proposed to find the better parameter value by using quantum-behaved particle swarm optimization(QPSO) which has better search ability. The QPSO-LSSVM model was trained and tested with polymyxin fermentation data-set. The results indicate that QPSO-LSSVM model can obtain better forecast effect than artificial neural network(ANN) model. Based on this model, QPSO algorithm was applied to optimize trajectories of pH and Do in each fermentation phase. The simulation results indicate that the method can improve the output of product.