佳木斯大学学报：自然科学版
佳木斯大學學報：自然科學版
가목사대학학보：자연과학판
Journal of Jiamusi University(Natural Science Edition)
2012年
1期
107-109
,共3页
姚尔果%闫秋粉%南振岐%薛小虎
姚爾果%閆鞦粉%南振岐%薛小虎
요이과%염추분%남진기%설소호
粒子群算法%进化速率%惯性权重因子%BP神经网络
粒子群算法%進化速率%慣性權重因子%BP神經網絡
입자군산법%진화속솔%관성권중인자%BP신경망락
particle swarm optimization algorithm%Evolutionary rate%Inertia weight factor%BP neuralnetwork
为解决BP神经网络局部性收敛度慢的问题，提出了基于改进粒子群算法的BP神经网络模型．该方法通过粒子群进化速率动态调整惯性权重因子，提高了算法的收敛速度和全局搜索最优值的能力．提出的模型和改进的算法模拟仿真表明：该方法对收敛速度和精度有更好的拟合性．
為解決BP神經網絡跼部性收斂度慢的問題，提齣瞭基于改進粒子群算法的BP神經網絡模型．該方法通過粒子群進化速率動態調整慣性權重因子，提高瞭算法的收斂速度和全跼搜索最優值的能力．提齣的模型和改進的算法模擬倣真錶明：該方法對收斂速度和精度有更好的擬閤性．
위해결BP신경망락국부성수렴도만적문제，제출료기우개진입자군산법적BP신경망락모형．해방법통과입자군진화속솔동태조정관성권중인자，제고료산법적수렴속도화전국수색최우치적능력．제출적모형화개진적산법모의방진표명：해방법대수렴속도화정도유경호적의합성．
To solve local convergence ot slow problems ot Bt＂ neural network, a tit＂ neural network model based on improved particle swarm optimization （PSO） algorithm was proposed. The convergence speed and theability of optimal value global searching were promoted by evolutionary rate adaptive inertia weight factor. Final- ly, the fitness of convergence speed and accuracy was validated by practical application through simulating the model and the improved algorithm.
