人工智能的原理及应用

介绍了人工智能的发展，并对改进的神经网络专家系统的优越性做了介绍，指出了神经网络专家系统广阔的应用前景及实现。     

人工神经网络理论及其应用

简要介绍了人工神经网络的发展过程和基本理论，从神经网络具有自学习功能、联想存储功能和高速寻找优化解的能力三个方面论述了其特点和优越性，详细阐述了在模式识别、信号处理、自动控制、人工智能、自适应的人机接口、优化计算、通信以及其它方面的应用，探讨和分析了其发展前景。     

Range of the Applicability of the Wilkins Method to the Investigation of the Dynamic Stress-Strain State of Anisotropic Elastic Axisymmetric Shells

The exactness and stability of the Wilkins method in the investigation of the stress-strain state of axisymmetric anisotropic elastic shells, including thick-walled ones, have been studied for different values of artificial viscosity, pulse load rise and fall time, and different deformation process durations. A comparison of numerical calculations, performed by means of an application software package developed by V. A. Romashchenko on the basis of the Wilkins method, with experimental results, known numerical data, and with calculations using the computation kernel LS-DYNA 3D has been carried out.     

Boole函数的线性可分性(Ⅰ)——n-维超立方体的基本理论

Boole函数的线性可分性是前向人工神经网络理论中的一个比较困难的问题之一。目前仅对变量数n≤7的某些问题给予讨论。当n≥8时,尚无判别Boole函数线性可分的一般准则,更无线性可分Boole函数的计数公式。基于此,本文详细地研究了与Boole函数线性可分性有关的n-维超立方体的基本理论,包括n-维超立方体的基本性质、超立方体中的平行线、子超立方体的计数等,并给出了构造n-维超立方体图的一种新方法。     

Effects of modeling means on properties of monitoring models of spot welding quality

Analyzing and modeling the relation between monitoring information during welding and quality information of the joints is the foundation of monitoring resistance spot welding quality. According to the means of modeling, the known models can be divided into three large categories: single linear regression models, multiple linear regression models and multiple non-linear models. By modeling the relations between dynamic resistance information and welding quality parameters with different means, this paper analyzes effects of modeling means on performances of monitoring models of resistance spot welding quality. From the test results, the following conclusions can be drawn: By comparison with two other kinds of models, artificial neural network (ANN) model can describe non-linear and high coupling relationship between monitoring information and quality information more reasonably, improve performance of monitoring model remarkably, and make the estimated values of welding quality parameters more accurate and reliable.     

储层敏感性快速预测软件在大港油田的研究与应用

在储层特征分析，室内实验，机理研究的基础上，利用人工神经BP网络建立预测模型，并编制了预测储层5种敏感性的软件，该软件受人为因素干扰小，所需参数少，准确度高，综合符合率大于80％现场应用结果表明，该技术能为制定保护油气层技术措施提供较可靠的依据，能提高油田滚动开发效益，改善老油田的开发效果，该软件对BP算法进行了改进，（1）从两方面入手使网络摆脱平坦区，一是对输入数据进行归一化处理，使O1的取值在[0,1]之间，二是一旦网络陷入平坦区域，局部极小，使连接权值Wkj,Wji和阈值Qk,Qj同时缩小一个因子，λ＞1，可使Ok(1-Ok)脱离零值，离开平坦区；（2）加速收敛，方法有自动调整学习因子，添加动量项以及对权值进行批处理，用BP算法预测储层潜在敏感性，首先应确定影响储层敏感性的主要因素，然后根据这些因素有针对性地收集有关资料并进行处理，再根据敏感性预测的要求，设计相应的网络结构进行训练，最后对训练好的网络进行.检验。     

根据人工裂缝特征优化油井压裂地质方案

近年来大庆油田外围地区地质储量得到动用，每年依靠压裂投产的井逐渐增多，由达西定律可知，在油层厚度一定的情况下，措施井的渗透率是影响产量的关键参数，油井压裂可以改善油井周围地层渗透性，消除近井地带油层污染，从而提高油井产能，而人工压裂的成功与否主要取决于裂缝的导流能力和裂缝的几何形态，为此，根据台105地区地质特征，结合低渗透油田开发实际，本着经济有效的开发原则，从人工裂缝特征，方位对开发效果的影响等方面，论述了优化油井压裂地质方案技术的方法，改善了低渗透油田的开发效果。     

Quantitative analysis of Ni effect on CCT diagrams of novel air-cooled bainite steels using artificial neural network models

The quantitative effect of Ni content on continuous cooling transformation （CCT） diagrams of novel air-cooled bainite steels was analyzed using artificial neural network models. The results showed that Ni may retard the high- and medium-temperature transformation and martensite transformation. The results conform to the materials science theories.     

Integration framework for intelligent manufacturing processes

In this article, we describe a new approach to applying distributed artificial intelligence techniques to manufacturing processes. The construction of intelligent systems is one of the most important techniques among artificial intelligence research. Our goal is to develop an integrated intelligent system for real time manufacturing processes. An integrated intelligent system is a large knowledge integration environment that consists of several symbolic reasoning systems (expert systems) and numerical computation packages. These software programs are controlled by a meta-system which manages the selection, operation and communication of these programs. A meta-system can be implemented in different language environments and applied to many disciplines. This new architecture can serve as a universal configuration to develop high performance intelligent systems for many complicated industrial applications in real world domains.To whom all correspondence should be addressed.     

Wavefront reduction using graphs,neural networks and genetic algorithm

This paper presents an algorithm for nodal numbering in order to obtain a small wavefront. Element clique graphs are employed as the mathematical models of finite element meshes. A priority function containing five vectors is used, which can be viewed as a generalization of Sloan's function. These vectors represent different connectivity properties of the graph models. Unlike Sloan's algorithm, which uses two fixed coefficients, here, five coefficients are employed, based on an evaluation by artificial neural networks. The networks weights are obtained using a simple genetic algorithm. Examples are included to illustrate the performance of the present hybrid method. Copyright © 2004 John Wiley & Sons, Ltd.