复杂系统近临界态分析与调控综述

Survey on Near-Critical Analysis and Control of Complex Systems

近年来,复杂系统近临界态的研究引起了诸多领域学者的广泛关注.近临界态是指复杂系统在连续相变点附近所处的状态.处于近临界态的复杂系统往往兼具鲁棒性和灵活性,能够涌现出有益的新特性,也可能在外界扰动下发生系统级大规模灾害.文章以复杂系统近临界态作为研究对象,首先描述了近临界态的性质及其对复杂系统的影响.随后基于当前的研究成果,详细分析了复杂系统近临界态的建模、预测以及调控相关的理论与方法.在建模方面,主要分析了元胞自动机、网络理论和相关计算、渗流模型、长瞬态以及随机噪声;在预测方面,详细阐述了基于时间序列和机器学习的预测方法;在调控方面,着重讨论了基于噪声和参数的调控理论与方法.文末总结并展望了复杂系统近临界态研究的发展趋势.

Recently,complex system’s near-critical state has received extensive attention of scholars in many fields.The near-critical state is the complex system’s state near the point of continuous phase transition.A complex system in the near-critical state often has the characteristics of robustness and flexibility,emerging positive new features.It may also suffer system-level large-scale disasters under external disturbances.This paper takes the near-critical state of complex systems as the research object.It firstly describes the near-critical state’s properties and its influence on complex systems.Then,based on the current research,it analyzes the theories and methods related to the near-critical state’s modeling,prediction and regulations in detail.In terms of modeling,cellular automata,network theory and related calculations,percolation models,long transients,and random noise are mainly discussed.As for prediction,prediction methods based on time series and machine learning are elaborated.In aspect of regulation,the control theory and method based on noise and parameters are emphatically analyzed.At last,it makes conclusions and provides prospects of the near-critical state’s future studies.