PSO-based Parameter Estimation of Nonlinear Kinetic Models for β-Mannanase Fermentation

Z. Liu, W. Qi, Z. He, H. Wang and Y. Feng

Abstract
Particle swarm optimization (PSO), as a novel evolutionary algorithm involved in social interaction for global space search, was firstly used in kinetic parameter estimation. Based on three developed nonlinear kinetic equations for bacterial cell growth, total sugar utilization and β-mannanase production by Bacillus licheniformis under the support of a batch fermentation process, various PSO algorithms as well as gene algorithms (GA) were developed to estimate kinetic parameters. The performance comparison among these algorithms indicates the improved PSO (Trelea 1) is most suitable for kinetic parameter estimation of β-mannanase fermentation. In order to find the physical-chemical-meanings of kinetic parameters from many optimized results, multiobjective optimization with a normalized weight method was adopted. The 9 desired parameters in equations were obtained by the Trelea 1 type PSO with two batches fermentation data, and the results predicted by the models were also in good agreement with the experimental observations.

This work is licensed under a Creative Commons Attribution 4.0 International License

Keywords
Particle swarm optimization, parameter estimation, kinetic model, multiobjective optimization, β-Mannanase