粒子群优化算法在非线性模型预测控制中的研究应用

Application of nonlinear model predictive control based on particle swarm optimization

提出了一种基于粒子群优化算法(PSO)的非线性模型预测控制(NMPC)。作为NMPC重要组成的滚动优化部分对控制效果的好坏起着关键的作用,因而寻求一种可靠的优化算法十分必要。PSO算法是一种群集智能方法,通过粒子之间的合作与竞争及进化实现对多维复杂空间的高效搜索,属于一类随机全局优化技术,已成功应用于各科学和工程领域。本文在滚动优化部分应用粒子群优化算法来求解预测控制律,对非线性系统施加优化控制,此外,对常规线性递减加权因子ω策略进行了讨论,提出了非线性递减策略,可进一步缩短优化时间和优化精度。仿真实验效果良好,验证了这种优化算法的正确性和有效性。

The objective of nonlinear model predictive control （NMPC） is to select a set of future control moves in order to minimize the function, based on a desired output trajectory over a prediction horizon. Thus it is neeessary to seek a highly efficient optimization method. A new NMPC based on particle swarm optimization （PSO） is presented here. The PSO algorithm employed exhibits good optimization performance. Furthermore the effect of different kinds of inertia weightω in the PSO is also diseussed in this paper. Satisfactory experimental results are obtained and demonstrate the efficacy of the PSO algorithm employed.