| 基于改进粒子群算法的烧结配料预测模型 |
| |
| 引用本文： | 黄晶晶,万丽丽,秦岭.基于改进粒子群算法的烧结配料预测模型[J].湖南农业大学学报(自然科学版),2011(3):90-94. |
| |
| 作者姓名： | 黄晶晶 万丽丽 秦岭 |
| |
| 作者单位： | （武汉工业学院 电气信息工程系,湖北 武汉430022） |
| |
| 摘 要： | 针对烧结配料系统中的非线性、复杂性和相关性，基于BP神经网络建立烧结配料的预测模型，并采用粒子群算法对预测模型参数进行优化。为了克服粒子群算法的局部收敛性，在迭代过程中，根据迭代次数对惯性权重进行动态非线性调整，从而提高算法的搜索能力。仿真结果表明，所提出的改进粒子群算法与传统的粒子群算法比较，收敛速度快、迭代次数少、具有较强的全局寻优能力。
|
| 关 键 词： | 配料优化 粒子群算法 BP神经网络 惯性权重 |
Prediction Model of Sintering Burden Based on Improved Particle Swarm Algorithm |
| |
| Institution: | (Electrical and Information Engineering Department, Wuhan Polytechnic University, Wuhan430023,China) |
| |
| Abstract: | The prediction modeling of the sintering experiment is proposed in this paper, based on BP neural network, according to the nonlinear, complexity and relativity of the sintering burden system. And the parameters of the prediction model are optimized using particle swarm algorithm. In order to overcome the local convergence of standard particle swarm algorithm, inertia weight in this paper is adjusted dynamically and nonlinearly according to the iteration times, so as to improve the algorithm''s searching capability. The simulating result shows that the improved particle swarm algorithm has faster converge, fewer iteration times and stronger global optimization ability, compared with other particle swarm algorithms. |
| |
| Keywords: | burdening optimization particle swarm algorithm BP neural network inertia weight |
|
| 点击此处可从《湖南农业大学学报(自然科学版)》浏览原始摘要信息 |
| 点击此处可从《湖南农业大学学报(自然科学版)》下载免费的PDF全文 |
