基于层次化可塑性回声状态网络的混沌时间序列预测  

Hierarchical plasticity echo state network for chaotic time series prediction

作　　者：那晓栋 王嘉宁 刘墨燃 任伟杰 韩敏[3,4] NA Xiao-dong;WANG Jia-ning;LIU Mo-ran;REN Wei-jie;HAN Min(Faculty of Electronic Information and Electrical Engineering,Dalian University of Technology,Dalian 116024,China;College of Intelligent Systems Science and Engineering,Harbin Engineering University,Harbin 150001,China;Key Laboratory of Intelligent Control and Optimization for Industrial Equipment of Ministry of Education,Dalian University of Technology,Dalian 116024,China;Professional Technology Innovation Center of Distributed Control for Industrial Equipment of Liaoning Province,Dalian University of Technology,Dalian 116024,China)

机构地区：[1]大连理工大学电子信息与电气工程学部,辽宁大连116024 [2]哈尔滨工程大学智能科学与工程学院,哈尔滨150001 [3]大连理工大学教育部工业设备智能控制与优化重点实验室,辽宁大连116024 [4]大连理工大学辽宁省工业装备分布式控制专业技术创新中心,辽宁大连116024

出　　处：《控制与决策》2023年第1期133-142,共10页Control and Decision

基　　金：国家自然科学基金项目(62173063);中央高校基本科研业务费专项资金项目(DUT20LAB114)。

摘　　要：为了提高回声状态网络对于混沌时间序列特征提取与预测的能力,提出一种层次化可塑性回声状态网络模型.该模型将多个储备池顺序连接,通过逐层特征变换的方式增强对非线性多尺度动态特征的提取能力.同时,引入神经科学中的内在可塑性机制模拟真实生物神经元的放电率分布,以最大化神经元的信息传递为目标对储备池进行预训练.层次化可塑性回声状态网络不仅能够增加模型的容量,降低随机投影所带来的不稳定性,而且也为理解储备池的表示、处理、记忆及储存操作提供一种新的思路.仿真实验结果表明,相比于其他7种改进的回声状态网络模型,所提出的模型在人造数据和真实数据所构成的混沌时间序列预测任务中均能取得最优的预测精度.To improve the ability of echo state network for feature extraction and prediction on chaotic time series,a hierarchical plasticity echo state network(HPESN)model is proposed.In this model,multiple reservoirs are connected in sequence,and the ability of nonlinear multi-scale dynamic feature extraction is enhanced through layer-by-layer feature transformation.Meanwhile,the intrinsic plasticity mechanism in neuroscience is introduced to simulate the firing rate distribution of real biological neurons,and the reservoir is pre-trained with the goal of maximizing neuronal information transmission.The HPESN not only increases the capacity of the model and reduces the instability caused by random projection,but also provides a new idea for understanding the representation,processing,memory and storage operations of the reservoir.The simulation results show that compared with other seven modified echo state network models,the proposed HPESN model achieves the best prediction accuracy in the chaotic time series prediction task composed of synthetic data and real-world data.

关 键 词：混沌时间序列预测 回声状态网络 层次化策略 神经内在可塑性 预训练 神经网络 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]