复杂系统的变结构分析

Structure change analysis of complex system

传统的变结构研究方法都是建立在模型基础上的 ,变结构分析主要集中在系统运行模型中参数的变化或变量的增减问题上 .这种传统的分析方法只限于人们对未知系统的模型结构有一个大致的了解 ,在一系列先验假设的条件下才能进行 .但是在系统建模之前 ,人们很难预先知道系统的运行的模型形式 .,对复杂系统而言 ,传统的变结构分析方法就难以奏效 .特别地 ,对非线性向量时间序列系统而言 ,系统内部动态均衡结构的变化不仅具有空间结构 ,而且具有一定的时间结构 .针对这种实际情况 ,本文提出了一种新的变结构分析思想 ,利用神经网络技术对系统的运行规律的结构变化情况进行了分析 .利用上海股市数据进行了实证研究 。

Traditional variable structure analysis depends on models. At present, structural change analysis focuses on the parameter's changes and variable's changes in quantitative models. These traditional analysis based on condition of researchers have acknowledge about the system rules and can provide the model forms. But in the actual systems, it is difficult to know the system rule before data analysis process, the traditional analysis method can not be suitable to complex systems. In order to solve this problem, we propose a new idea to detecting structure changes in complex system. We apply neural network technique to analyze system running rules, establish two kinds of algorithms for detecting structural change in actual system. As an empirical case, we discuss the Shanghai Stock Market using our new algorithm, and have got satisfied result.