摘要
将一种基于混沌退火的Hopfield神经网络(ACHN)应用到储层聚类分析中。这里主要阐述了ACHN聚类的基本原理、混沌神经元的标定、ACHN聚类的策略、先验信息的约束以及混沌退火搜索。实际砂砾岩储层聚类分析表明:这种网络能克服普通Hopfield网络陷入局部最优的不足,得到更准确、精细的聚类结果。
In this paper,an annealed chaotic Hopfield network(ACHN) is proposed and applied in clustering of the gravel body reservoir.In the paper,the basic theory and strategy of ACHN clustering are discussed in detail and how to mark chaotic neurons,constrain Hopfield network with known information,and anneal with chaos is also introduced.The much fine clustering results show that the method can successfully solve the problem of local minima in Hopfield network,and obtain the global optimal solution.
出处
《物探化探计算技术》
CAS
CSCD
2006年第2期89-92,共4页
Computing Techniques For Geophysical and Geochemical Exploration
