摘要
为了提高人工鱼群算法AFSA(artificial fish swarm algorithm)的全局搜索能力及加快其收敛速度,提出一种将其与免疫算法IA(immune algorithm)进行结合的新方法,形成了免疫人工鱼群算法IAFSA(immuneartificial fish swarm algorithm),并且利用该算法自动选取径向基函数RBF(radial basis function)神经网络中的输入变量,以及对网络中隐含层到输出层之间的权值进行训练,从而减少了RBF神经网络的工作量,提高了训练速度。用优化后的RBF神经网络进行短期负荷预测,结果表明,该方法具有较高的预测精度。
In order to enhance global search capability and accelerate convergence rate of artificial fish swarm algorithm（AFSA）,a new method is presented in which AFSA is combined with immune algorithm（IA）,thus forming immune artificial fish swarm algorithm（IAFSA）.The IAFSA is used to choose radial basis function（RBF） neural network＇s input variables automatically and train the weights that between network＇s hidden layer and output layer,which reduce RBF neural network＇s workload and improve it＇s training speed.Using the optimized RBF neural network to do short-term load forecasting,the results show that the proposed method has higher forecasting precision.
出处
《电力系统及其自动化学报》
CSCD
北大核心
2011年第2期142-146,共5页
Proceedings of the CSU-EPSA
关键词
负荷预测
神经网络
人工鱼群算法
免疫算法
输入变量选择
径向基函数
load forecasting
neural network
artificial fish swarm algorithm（AFSA）
immune algorithm
input variables selection
radial basis function（RBF）
