非线性时间序列高性能仿真算法研究  

Research on Nonlinear Time Series' High Performance Simulation Algorithms

作　　者：邵晨曦[1,2] 童松桃[2] 杨明[1] 王子才[1] 

机构地区：[1]哈尔滨工业大学控制与仿真中心,哈尔滨150001 [2]中国科学技术大学计算机科学与技术系,合肥230027

出　　处：《系统仿真学报》2009年第15期4598-4602,共5页Journal of System Simulation

基　　金：国家自然科学基金(60874065;60434010)

摘　　要：系统相空间重构和几何特征值计算是非线性动力系统仿真的重要过程。因其算法时空复杂度较高而计算耗时,为实时仿真非线性时间序列带来了困难。对相图重构算法和几何特征值计算算法进行了优化,使其时间复杂度有不同程度的降低;进而并行化这些算法,使其效率进一步提升,最后算法运行时间缩短为原运行时间的1/30。以长程睡眠脑电数据分析和仿真为例,分析结果表明关联维的高低能够反映出睡眠的阶段,据此对非快速眼动睡眠期的仿真预测与实际结果相符。Reconstitution of phase graphics and computation of geometrical eigenvalues are important processes in nonlinear system＇s simulation. It＇s difficult to simulate nonlinear system such as EEG real-time because the algorithms of phase graphics reconstitution and geometrical eigenvalue computation are complex on both time and space. The algorithms were optimized to reduce their complexity, then the algorithms were parallelized, at last the integrated algorithm＇s running time was 1/30 of the running time before optimization and parallelization. The value of correlation dimension can reflect sleep stages after analyzing and simulating the sleep EEG, and then simulation forecast for NREM sleep stage was made which accorded with actual result.

关 键 词：非线性时间序列 仿真算法 算法优化 并行算法 睡眠脑电 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]