【摘要】 粒子群优化(PSO,Particle swarm optimizer)算法是基于群智能的全局优化技术,它通过粒子间的相互作用,对解空间进行智能搜索,从而发现最优解。其优势在于操作简单,容易实现,并且功能强大。目前PSO已成为国际演化计算界研究的热点。本文将PSO与模糊优选人工神经网络进行融合,对流域年径流量进行智能预测,在对模糊优选神经网络训练中采取PSO算法和梯度下降算法相结合的方法,充分发挥PSO全局寻优的能力和梯度下降局部细致搜索优势,结果表明采用这种方法可以提高模糊优选人工神经网络的训练效率并且具有很好的推广能力。更多还原

【Abstract】 PSO is a kind of stochastic global optimizer based on swarm intelligence, through the interaction between particles, PSO searches the solution space intelligently and find out the best. The advantage of PSO is that it is easy to operate and powerful. At present, PSO has become the research hotspot all around the world. A model integrating PSO and FNN (fuzzy artificial neural network) is established in this paper, and runoff prediction case was tested, which took full use of the global optimization of PSO and local accurate searching of BP. The case of runoff prediction shows that PSO-FNN is more efficient and has good generalization.更多还原