摘要
为了提高基本粒子群优化(PSO)算法的收敛性,提出了一种引入选择与变异机制的改进PSO算法。该算法选择一定范围的优秀粒子代替较差粒子,并使粒子以不同的概率变异。仿真试验表明,引入选择与变异机制使PSO算法的收敛速度得到了提高,并且有效抑制了PSO算法的早熟。将改进PSO算法应用于热工过程模型辨识,在较少的迭代次数内得到了比较精确的模型辨识结果,且具有很好的收敛性能,获得了满意的辨识效果。
In order to enhance the convergent behavior of the basic particle swarm optimization （PSO） algorithm,an improved PSO algorithm, into which the selection and mutation mechnisms being introduced,has been put forward. In the improved algorithm, a range of excellent particles is selected to substute the poor particles,and make the particles to mutate with different probability. Emulation test shows that the introduction of selection and mutation mechnisms makes the covergent rate PSO algorithm to be enhanced, and the precocity of PSO algorithm to be effectively restrained. The improved PSO algorithm has been used in identification through thermodynamic process model, a more precise result of model identification can be achieved in a smaller number of iterations, boasting very good convergent behavior,obtaining satisfactory results in identificarion.
出处
《热力发电》
CAS
北大核心
2010年第3期97-100,103,共5页
Thermal Power Generation
关键词
PSO算法
选择与变异
热工过程
模型辨识
收敛性
PSO algorithm
selection and mutation
thermodynamic process
model identification
