Parallel evolutionary modeling for nonlinear ordinary differential equations

We introduce a new parallel evolutionary algorithm in modeling dynamic systems by nonlinear higher-order ordinary differential equations (NHODEs). The NHODEs models are much more universal than the traditional linear models. In order to accelerate the modeling process, we propose and realize a parallel evolutionary algorithm using distributed CORBA object on the heterogeneous networking. Some numerical experiments show that the new algorithm is feasible and efficient. Foundation item: Supported by the National Natural Science Foundation of China (No. 70071042 and No. 60073043) Biography: Kang Zhuo (1970-), male, Lecturer, research interest: network computing and evolutionary computation.     

Parallel implementations of modeling

First, an asynchronous distributed parallel evolutionary modeling algorithm (PEMA) for building the model of system of ordinary differential equations for dynamical systems is proposed in this paper. Then a series of parallel experiments have been conducted to systematically test the influence of some important parallel control parameters on the performance of the algorithm. A lot of experimental results are obtained and we make some analysis and explanations to them. Foundation item: Supported by the National Natural Science Foundation of China (60133010, 70071042, 60073043) Biography: Cao Hong-qing ( 1972-), female, Associate professor, research direction; evolutionary computing, parallel computing.     

用遗传算法建立连续系统低阶近似模型的方法及其在电机控制中的应用

基于高阶非线性系统的阶跃响应数据，用遗传算法寻找最优参数，直接得到便于工程设计和分析的低阶线性连续系统模型．首先讨论了遗传算法及其改进方法，然后分析低阶系统参数变化范围的自动修改规则和建模方法，最后建立了交流电机调速系统的低阶近似模型．     

Genetic programming-based chaotic time series modeling

This paper proposes a Genetic Programming-Based Modeling (GPM) algorithm on chaotic time series. GP is used here to search for appropriate model structures in function space, and the Particle Swarm Optimization (PSO) algorithm is used for Nonlinear Parameter Estimation (NPE) of dynamic model structures. In addition, GPM integrates the results of Nonlinear Time Series Analysis (NTSA) to adjust the parameters and takes them as the criteria of established models. Experiments showed the effectiveness of such improvements on chaotic time series modeling.     

SW—GP组合算法在煤矿监测数据预警预报中的应用

为了更有效地利用煤矿监测数据进行煤矿安全事故的预警预报,提出滑动窗口-遗传程序设计组合算法实现了监测数据的动态预测。在程序设计时,数据采样采用滑动窗口技术实现,通过遗传程序设计算法实现系统的自适应建模。通过对监测数据的测试,证明了组合算法建立模型的预测值和实际结果具有很好的一致性。     

基于结构的神经网络在液压缓冲器建模中的应用

分析了液压缓冲器的结构及其动态工作过程，介绍了基于结构的神经网络建模方法．该建模方法根据系统结构组成特点将复杂系统分解为相互关联的简单子系统，用函数链神经元分别建立子系统模型，然后根据子系统间固有的连接关系将子系统神经元模型连接成一个网络，所得网络模型即为原系统模型．应用该方法建立了52SFZ—140—207B液压缓冲器的动态模型．结果表明，基于结构的神经网络建模方法对复杂非线性系统建模是有效的．     

贝叶斯型的时变系数回归模型

提出了具有时变系数的回归模型及其模型识别和参数估计的贝叶斯方 法．在此模型中对于时变回归系数的变动，应用了高斯型概率差分方程式作 为约束条件，称之为高斯型平滑性事先分布．模型中的超参数（ｈｙｐｅｒｐａｒａｍｅ－ ｔｅｒ）的估计，采用了最大似然估计法．模型的识别（差分次数的决定）应用了 Ａｋａｉｋｅ的最小 ＡＢＩＣ法．给出了模型估计的算法及其应用例子．最后，讨论了 平滑性事先分布中参数的最优估计的意义．     

Time-Based Dynamic Trust Model Using Ant Colony Algorithm

The trust in distributed environment is uncertain, which is variation for various factors. This paper introduces TDTM, a model for time-based dynamic trust. Every entity in the distribute environment is endowed with a trust-vector, which figures the trust intensity between this entity and the others. The trust intensity is dynamic due to the time and the inter-operation between two entities, a method is proposed to quantify this change based on the mind of ant colony algorithm and then an algorithm for the transfer of trust relation is also proposed. Furthermore, this paper analyses the influence to the trust intensity among all entities that is aroused by the change of trust intensity between the two entities, and presents an algorithm to resolve the problem. Finally, we show the process of the trusts＇ change that is aroused by the time＇s lapse and the inter-operation through an instance.     

基于径向基函数神经网络的热工过程模型辨识

为了准确反映热工过程动态特性,实现热工过程整体优化控制,提出了一类新的径向基函数神经网络(RBF-NN)的建模方法:采用熵方法和竞争学习算法,结合非线性自回归滑动平均(NARMA)模型的输入/输出结构实现RBF-NN的优化,辨识RBF-NN结构,并用最小二乘算法(LS)确定权向量,实现了典型的非线性热工过程建模。通过两个实例验证:基于NARMA结构的RBF-NN建模,具有较高的辨识精度和较少的隐层节点。     

动态模糊神经网络研究

针对静态网络无法处理暂态问题，对具有递归环节的动态模糊神经网络进行了研究。通过在网络第二层中加入内部反馈连接，使其具有动态映射能力，从而对动态系统有更好的响应。网络使用遗传算法与反向传播BP(BackPropagation)算法相结合来训练，避免陷入局部最优解。采用时序预测和动态非线性系统进行了仿真研究，结果表明，动态模糊神经网络较之普通模糊神经网络在收敛速度、预测精度和网络规模等方面都有较大的改善，并具有更好的动态系统处理能力。     

动态系统的演化建模

针对传统方法解决动态系统微分方程建模问题所遇到的困难和存在的不足,设计将方程进行串结构编码并用进化方法进行演化建模的算法,以串形结构表示结构,用进化算法优化结构和参数,成功地实现了动态系统的常微分方程组建模过程的自动化。计算实例表明:采用此算法能够在极短的时间内由计算机自动发现多个较优的常微分方程组模型,与原来GA和GP结合的方法相比较,它具有建模过程智能化、模型结构非常灵活多样、数据拟合和预测精度更高等优点。     

基于神经网络的非线性前馈补偿广义预测自校正控制器

采用多层前馈网络结构进行动态建模,并用Ｄａｖｉｄｏｎ最小二乘法作为在线学习算法,将辨识后得到的模型进行线性化．基于线性化模型设计广义预测控制器。将其与非线性前馈相结合,建立了一种适合于非线性系统的前馈补偿广义预测自校正控制器．仿真结果验证了本控制器对非线性系统控制的有效性     

Remote intelligent identification system of structural damage

The focus of this paper is to build the damage identify system, which performs “system identification“ to detcct the positions and extents of structural damages. The identification of structural damage can be characterized as a nonlinear process which linear prediction models such as linear regression are not suitable. However. neural network techniques may provide an effective tool for system identification. The method of damage identification using the radial basis function neural network (P, BFNN) is presented in this paper. Using this method, a simple reinforced concrete structure has been tested both in the absence and presence of noise. The resuits show that the RBFNN identification technology can he used with related success for the solution of dynamic damage identification problems, even in the presence of a noisy identify data. Furthermore, a remote identification system based on that is set up with Java Technologies.     

混合自回归模型的谱分析

线性时间序列模型谱密度的计算可以直接由定义获得,而非线性时间序列模型谱密度的计算目前还没有一般的理论.已有研究者将AR模型推广到MAR(混合自回归)模型,并且讨论了该模型的参数估计及模型选择问题.作者利用全期望公式及差分方程理论研究了混合自回归时间序列模型的谱分析,导出了自协方差函数的递推公式,给出计算谱密度的算法,并对一些常见的特殊情形给出了谱密度的具体表达式.     

Dynamic prediction of building subsidence deformation with data-based mechanistic self-memory model

This paper describes a building subsidence deformation prediction model with the self-memorization principle.According to the non-linear specificity and monotonic growth characteristics of the time series of building subsidence deformation,a data-based mechanistic self-memory model considering randomness and dynamic features of building subsidence deformation is established based on the dynamic data retrieved method and the self-memorization equation.This model first deduces the differential equation of the building subsidence deformation system using the dynamic retrieved method,which treats the monitored time series data as particular solutions of the nonlinear dynamic system.Then,the differential equation is evolved into a difference-integral equation by the self-memory function to establish the self-memory model of dynamic system for predicting nonlinear building subsidence deformation.As the memory coefficients of the proposed model are calculated with historical data,which contain useful information for the prediction and overcome the shortcomings of the average prediction,the model can predict extreme values of a system and provide higher fitting precision and prediction accuracy than deterministic or random statistical prediction methods.The model was applied to subsidence deformation prediction of a building in Xi’an.It was shown that the model is valid and feasible in predicting building subsidence deformation with good accuracy.     

Multi-Path Routing and Resource Allocation in Active Network

An algorithm of traffic distribution called active multl-path routing (AMR) in active network is proposed. AMR adopts multi path routing and applies nonlinear optimize approximate method to distribute network traffic among muhiple paths. It is combined to bandwidtb resource allocation and the congestion restraint mechanism to avoid congestion happening and worsen. So network performance can be improved greatly. The frame of AMR includes adaptive traffic allocation model, the conception of supply bandwidth and its‘ allocation model, the principle of congcstion restraint and its‘ model, and the implement of AMR based on multi-agents system in active network. Through simulations. AMR has distinct effects on network performance. The results show AMR is a valid traffic regulation algorithm.     

The dynamic modeling algorithm of information organization and allocation problem on internet communications

With the frequent information accesses from users to the Internet, it is important to organize and allocate information resources properly on different web servers. This paper considers the following problem: Due to the capacity limitation of each single web server, it is impossible to put all information resources on one web server. Hence it is an important problem to put them on several different servers such as: (1) the amount of information resources assigned on any server is less than its capacity; (2) the access bottleneck can be avoided. In order to solve the problem in which the access frequency is variable. this paper proposes a dynamic optimal modeling. Based on the computational complexity results, the paper further focuses on the genetic algorithm for solving the dynamic problem. Finally we give the simulation results and conclusions. Foundation item: Supported by the Hi-tech Research and Development Program of China(2002AAlZ1490) Biography: Li Yuan-xiang( 1963-), male, Prof, research direction; parallel computing,evolutionary hardware.     

Rosenau-Burgers方程的一个新的差分方法

从动力学系统的实际问题出发,针对Rosenau-Burgers方程的初边值问题进行了数值研究,揭示了复杂离散动态系统理论中非线性波耗散问题. 在方程求解的时间和空间区域,采用网格化方法,提出了一个新的三层隐式差分格式,对差分解进行了先验估计,并给出了该格式的稳定性和收敛性的严格理论证明. 数值实验的结果表明,差分格式简单而有效、计算速度快、稳定性好,并且差分格式使用了加权方法,使其具有普遍意义和推广价值.     

含饱和特性的双采样率数据Hammerstein系统辨识

针对含有饱和特性的双采样率数据Hammerstein系统提出了一种新的辨识方法。首先,将含有饱和特性的静态非线性环节和线性动态环节的串联,整理成一个非线性基函数和线性动态环节的串联。在此基础上,利用辅助模型辨识原理解决数据缺失、中间未知变量、被辨识参数之间存在耦合的问题, 通过递推辨识算法利用双率采样数据辨识单率Hammerstein模型中的参数。最后,以一个含饱和特性非线性系统实例的建模来验证提出辨识算法的有效性。     

基于IRBFNN和IRAN的非线性动态系统在线自适应建模

将资源分配网络算法(RAN)与相似隐单元合并操作、冗余隐单元删除操作和基于滑动数据窗连接权值学习相结合,形成了改进的资源分配网络(IRAN)算法。IRAN算法用于非线性动态系统的在线建模,能有效地改善模型精度和泛化能力。将改进径向基函数(RBF)神经网络(IRBFNN)和IRAN结合可以用于不确定非线性动态系统自适应建模。仿真研究表明:所提出的建模方法在模型精简、泛化和自适应等方面均具有优良的性能。