基于RETE算法的速生丰产林培育知识系统的设计

本文应用实现产生式系统的高效模式匹配算法—RETE算法,分析设计了速生丰产林生长培育知识系统。该系统的特点是使用相似性方法来处理模糊问题,能够较为理想地实现速生丰产林生长培育过程中的推理机制。     

Application of Artificial Neural Network in Elastic Internal Force Analysis of Two Way Slab

In this paper, the Artificial Neural Network (ANN) is applied to the internal force analysis of structure. After introducing the model and algorithm of back-propagation network, a trained BP network of three layers is used to calculate the maximum of elastic moment of middle span of two-way slab. In order to enhance the generalized capability of network, the modified value of weight should be regarded as convergence standard and in order to accelerate the learning process without vibration, the method of adding momentum coefficient is adopted. The analysis program of BP network is carried out with the software of Matlab. The result demonstrates that application of ANN in structure analysis is feasible.     

Improvement and Application on Back Propagation Network Based on Partial Least-squares Algorithm

This paper proposes a novel BP network model based on nonlinear iterative partial least-squares algorithm which can fit nonlinear data. The novel BP network model can reduce iterative step number and advance learning effieieney. This paper pretreats data by nonlinear iterative partial least-squares algorithms. The weights initialization of input floor and output floor are set by applying the loading weights of dependent variable and cause variable, the member of hidden nodes are set by applying factor numbers of nonlinear iterative partial leastsquares algorithm, the connection co- efficient is set by applying the connection matrix B. Performances of the BP, PLS, and PLS-BP are analyzed and compared. The results show that the PLS-BP has better fitting and forecasting than BP and PLS.     

关于近交系数计算方法的改进

本文对经典方法予以较大的改进.特别是对于有规世代随机出现的系谱中的近交系数,本文给出分段迭代算法,继而大大地提高了近交系数的计算速度.此外,对于多始祖大系谱,我们给出分解算法,使得查找通径和计算工作极易进行.     

Boundary Element Methods with Overlapping Domain Decomposition for Parabolic Problem

Boundary element method is a numerical method for solving partial differential equations. There are several formulations of boundary element method (BEM) applied to solve a parabolic differential equation.The approach,which employs time- dependent fundamental solution,allows longer time steps in time integration than other approaches,and this can cut down on time for computer implementation with high precision.Domain decomposition method,which decompose the domain that a given problem is to be solved into subdomains,has the advantages of reducing the large problem into smaller ones and reducing the complex problem into simpler ones,and allows parallel computing.An overlapping domain decomposition method is applied combining a boundary element formulation with time-dependent fundamental solution to solve a diffusion equation. Firstly, by domain decomposition, the problem divided into two problems on subdomains, and then the initial-Boundary problems are solved by boundry element method on each subdomain.Some numerical examples are presented to illustrate feasibility and efficiency of the method. The numerical experiments show that the convergence rate of the method is dependent with the overlapping degree of the subdomains.     

Using a Neural Network with Nonlinear Self-feedback to Solve the Maximum Clique Problem

As the MCP is NP-hard, an efficient approach to treating this problem is to design appropriate recurrent neural networks. We develop a new algorithm for the MCP, which can, to a certain extent, prevent the associated neural network from falling into local optimal points. The proposed algorithm incorporates nonlinear self-feedback into the SLDN algorithm and has distinguished dynamical characteristics. Simulation results show that the performance of proposed algorithm is statistically superior to the SLDN algorithm.     

Choosing Method of Wavelet Neuron Based on ECG Data Compression

The paper discusses a wavelet network for the ECG data compression and proposes the method for choosing its wavelet neuron.According to the spectrum range of the ECG data,we decide the time-frequency field of ECG.And the time-frequency field of wavelet is also determined by the spectrum range of it.The wavelet neuron is fixed preliminarily by the first two steps.Then the preliminary wavelet neuron is screened by using OLS algorithm.We choose Morlet as the mother wavelet,and use the ECG signal to validate by the method.The result demonstrates that the number of Morlet whose spectrums locate at the ECG's is up to 152.But after screening by the OLS algorithm,it reduces sharply.This method can make the size of the wavelet network driving to optimum and also reduce the training time of the wavelet network sharply.     

Learning Causality Diagram Structure Based on Genetic Algorithm

The Causality Diagram theory,which adopted graphical expression of knowledge and direct causality intensity of causality,overcomes some shortages in Belief Network and has evolved into a mixed causality diagram methodology coped with discrete and continuous variable.But it is difficult that the structure of Causality Diagram given by expert.Because the complexity of causality diagram structure goes up exponentially through the number of the vertex's increasing,it is NP-hard problem to find the most possible structure from a set of data.The authors discuss approaches and present Genetic Algorithm,to find the most possible structure from a set of data.Experiment shows the method is effective.     

Fault Detection Techniques for Feeder Automation

Fault detection of the distribution systems is one of the key techniques for feeder automation. At all times, it is still very difficult for the exact detection of single phase to ground fault. First of all, the distinguishing rules of the faulted phase are presented when ground fault occurs in distribution systems. This paper analyzes the fault detection technique for single phase to ground in system with floating neutral point. Based on faulted phase's current, a method realizing ground fault protection is proposed and the algorithm model of action criterion is given out. The simulation results show that the method will be with very good selectivity to ground fault, and it can be used preferably in feeder automation. The detection method for short circuit fault is also introduced simply.