摘要
提出一种ε机理论发现复杂系统中隐含模式的新思路。该方法能在对系统内部结构无法知晓的情况下,不需要假设而根据系统自身的特性建模。先对时间序列进行符号化处理,生成符号序列。再采取因果态分割重构法重构ε机,重构后能够显示出原始序列中的隐含信息。通过对工程中配电网系统故障选线的多组实验证明,4个评判指标:熵率、统计复杂度、状态数和相对熵率能够一致地显示出各条线路的隐含信息,在几个关键影响因素改变的情况下均能有效地区分故障线路和非故障线路。从而验证了ε机理论在复杂系统建模中的优越性,为复杂系统建模提供了一种有效途径,也为配电网故障选线提供了一种新的方法。
A new method using ε-machine theory to find hidden pattern of complex system is proposed.This method could build model based on the characteristics of the system itself without assuming or knowing the internal structure of the system.First,the symbolic pro-cessing is performed on the time series to construct the symbol series.The method of casual state splitting reconstruction（CSSR）is then adopted to reconstruct theε-machine,which could show the hidden pattern of the original series.Several experiments on fault line selec-tion of a power distribution system prove that the four evaluation indexes：entropy rate,statistic complexity,state number and relative en-tropy rate can show the hidden pattern of each line consistently,and distinguish fault lines and non-fault lines effectively in the situation where several key factors change.Thus the superiority ofε-machine theory in complex system modeling is verified,which provides an ef-fective way for complex system modeling,and also provides a new method for fault line selection in power distribution network.
出处
《仪器仪表学报》
EI
CAS
CSCD
北大核心
2014年第8期1758-1765,共8页
Chinese Journal of Scientific Instrument
基金
国家自然科学基金(61077071)
河北省自然科学基金(F20110203207)
河北省高等科学技术研究重点项目(ZD2014100)资助项目
