摘要
为了提高网络大数据计算的速度和收敛性,针对当前仿生群算法容易出现局部收敛的问题,提出一种基于优化粒子群智能信息处理的网络大数据计算方法。采用特征尺度标识方法进行网络计算数据集规范化处理,结合粒子群算法进行网络计算的大数据聚类分析,根据粒子速度和位置更新迭代公式确定初始聚类中心,通过混沌差分扰动进行个体寻优,降低群体适应度方差,使得计算程序满足收敛法则,提高网络计算的效率。仿真结果表明,采用该方法进行网络大数据的智能计算能得到最优适应度值,收敛性和处理速度都具有优势。
In order to improve the speed and convergence of the network big data calculation, and solve the problem that the bionic swarm algorithm is prone to appear the local convergence, a network big data computing method based on optimized particle swarm intelligent information processing is put forward. The characteristic scale identification method is used to perform the dataset standardization processing of network computing, and combined with the particle swarm algorithm to conduct the big data clustering analysis of network computing. According to the update iterative formula of the particle velocity and position, the initial clustering center is determined. The individual is optimized by means of chaotic difference disturbance to reduce the fitness variance of population, which makes that the calculation program can meet the convergence rule, and the efficiency of network computing improved. The simulation results show that the method can obtain the optimal fitness value, fast convergence and processing speed while performing the intelligent computing of the network big data.
出处
《现代电子技术》
北大核心
2017年第15期41-43,共3页
Modern Electronics Technique
关键词
智能信息处理
粒子群
网络计算
大数据
intelligent information processing
particle swarm
network computing
big data
