一种改进细菌觅食优化算法及其在软测量建模中的应用

摘要

Demand for model parameter optimization in soft measurement modeling, on the basis of analyzing bacteria foraging optimization algorithm ( BFOA) and particle swarm optimization ( PSO ) algorithm, a novel bacterial foraging particle swarm based hybrid optimization algorithm ( BSOA) is proposed by taking advantage of bolh BFOA and PSO. The new algorithm introduces particle moving inspiration of PSO into BFOA, which effectively solves the blindness of the location update in BFOA. The new method is used for typical function optimization and optimization of the parameters of least squares support vector machine(LSSVM) model in research octane number ( RON ). Simulation results show that this method enhances the global optimization capability and convergence rate of the algorithm, to some extent, improves the prediction precision and generalization ability of the model too.%针对软测量建模中模型参数的优化需求,在分析细菌觅食优化算法(BFOA)和粒子群优化(PSC)算法的基础上,将二者有机结合,提出了一种新型细菌觅食粒子群混合优化算法(BSOA).该算法将PSO粒子移动的思想引入BFOA,有效解决了BFOA趋向性操作中细菌位置更新的盲目性.将其分别用于典型函数的寻优与成品油研究法辛烷值最小二乘支持向量机(LSSVM)模型参数的优化,仿真结果表明:该方法有效增强了算法的全局寻优能力与收敛速度,并在一定程度上改善了模型的预测精度与泛化能力.