Learning Automata from Ordered Examples

Connectionist learning models have had considerable empirical success, but it is hard to characterize exactly what they learn. The learning of finite-state languages (FSL) from example strings is a domain which has been extensively studied and might provide an opportunity to help understand connectionist learning. A major problem is that traditional FSL learning assumes the storage of all examples and thus violates connectionist principles. This paper presents a provably correct algorithm for inferring any minimum-state deterministic finite-state automata (FSA) from a complete ordered sample using limited total storage and without storing example strings. The algorithm is an iterative strategy that uses at each stage a current encoding of the data considered so far, and one single sample string. One of the crucial advantages of our algorithm is that the total amount of space used in the course of learning for encoding any finite prefix of the sample is polynomial in the size of the inferred minimum state deterministic FSA. The algorithm is also relatively efficient in time and has been implemented. More importantly, there is a connectionist version of the algorithm that preserves these properties. The connectionist version requires much more structure than the usual models and has been implemented using the Rochester Connectionist Simulator. We also show that no machine with finite working storage can iteratively identify the FSL from arbitrary presentations.     

Characterization of EBPVD micro-layer composites and simulation of its internal stress state

Based on the basic operating principal and the technology characteristic of electron beam physical vapor deposition(EBPVD) technique, EBPVD was used to prepare the micro-layer composites. The effect on the substrate preheating temperature was taken into accounts and the finite element analysis package ANSYS was used to simulate the internal stress field and the potential displacement changing tendency. The results show that one of the most important quality factors on the judgment of micro-layer composites is the adhesion between the substrate and the deposition layers as well as among the different deposition layers. Besides the existance of temperature gradient through the thickness of layers, the main reason for the internal stress in micro-layer composites is the mismatch of various properties of the layer and the substrate of different thermal expansions and crystal lattice types. With the increase of substrate preheating temperature, the inter-laminar shear stress also takes on a tendency of increase but the axial residual stress decrease.     

正则化长波方程的三次样条差分模拟

本文对正则化长波方程(RLW方程)的时间、空间项分别采用三次样条公式进行离散,所得到的差分方程为三点、三层隐格式,对时间的精度为四阶,对空间的精度为二阶,每个时间步只需求解一个三对角方程组,而三次样条配置法在每个时间步中需要求解三个三对角方程组。本文对单个孤立波的演化进行了计算,参数与文[8]相同,其结果与解析解的最大绝对误差均保持在10~(-3)的量级以内,与文[8]给出的数值方法精度相当。对两、三个孤立波的相互碰撞,本文结果显示出在忽略10~(-3)量级的误差后,碰撞前后的孤立波形状保持不变,无论是一个孤立波的演化,还是两、三个孤立波相互碰撞,当将波形放大100倍后均可以发现有振荡的尾波,这说明这类振荡尾波的产生并非真实的物理现象,而是由数值误差产生的。本文认为仅从工程应用的意义上说,RLW方程具有孤立子性质的结论是可以接受的。     

Experimental and numerical investigation of fracture in double-edge notched steel plates

Double-edge notched (DENT) steel plates were pulled until complete fracture and several experimental observations were made (using profilometry and scanning electron microscopy). The essential work of fracture (EWF) model was found to be well verified. Numerical simulations—up to the maximum load only—of some experiments were performed using the finite element method (FEM), and incorporating geometric and material non-linearities (large deformation elasto-plasticity). Some experimental measurements were compared with the corresponding numerical computations and excellent agreement was found.     

Analysis of brushless three-phase synchronous generator without exciter

Recently, a demand for small-capacity generators has been increasing as electric sources in small ships and automobiles or as portable electric sources driven by engines. It is desired that the structure of small-capacity generators be simple and robust, and that the generators be highly reliable, easily maintained and controlled. This paper describes an analysis of the original brushless synchronous generator without exciter. The output voltage can be adjusted in the wide range by controlling the stator dc current. To analyze the characteristics of this generator, the finite element method is applied. It is found that the results of theoretical analysis agree well with the experimental results.     

Least-squares integration of one-dimensional codistributions with application to approximate feedback linearization

We study the problem of approximating one-dimensional nonintegrable codistributions by integrable ones and apply the resulting approximations to approximate feedback linearization of single-input systems. The approach derived in this paper allows a linearizable nonlinear system to be found that is close to the given system in a least-squares (L ₂) sense. A linearly controllable single-input affine nonlinear system is feedback linearizable if and only if its characteristic distribution is involutive (hence integrable) or, equivalently, any characteristic one-form (a one-form that annihilates the characteristic distribution) is integrable. We study the problem of finding (least-squares approximate) integrating factors that make a fixed characteristic one-form close to being exact in anL ₂ sense. A given one-form can be decomposed into exact and inexact parts using the Hodge decomposition. We derive an upper bound on the size of the inexact part of a scaled characteristic one-form and show that a least-squares integrating factor provides the minimum value for this upper bound. We also consider higher-order approximate integrating factors that scale a nonintegrable one-form in a way that the scaled form is closer to being integrable inL ₂ together with some derivatives and derive similar bounds for the inexact part. This allows a linearizable nonlinear system that is close to the given system in a least-squares (L ₂) sense together with some derivatives to be found. The Sobolev embedding techniques allow us to obtain an upper bound on the uniform (L _∞) distance between the nonlinear system and its linearizable approximation. This research was supported in part by NSF under Grant PYI ECS-9396296, by AFOSR under Grant AFOSR F49620-94-1-0183, and by a grant from the Hughes Aircraft Company.     

NUMERICAL SIMULATIONS OF WHOLE FIELD BEHAVIOR FOR ROUND BUOYANT JET

ＮＵＭＥＲＩＣＡＬＳＩＭＵＬＡＴＩＯＮＳＯＦＷＨＯＬＥＦＩＥＬＤＢＥＨＡＶＩＯＲＦＯＲＲＯＵＮＤＢＵＯＹＡＮＴＪＥＴＬｉＷｅｉ，ＨｕａｉＷｅｎ－ｘｉｎ（ＷｕｈａｎＵｎｉｖｅｒｓｉｔｙｏｆＨｙｄｒａｕｌｉｃａｎｄＥｌｅｃｔｒｉｃＥｎｇｉｎｅｅｒｉｎｇ，...     

An eight‐node hybrid‐stress solid‐shell element for geometric non‐linear analysis of elastic shells

This paper presents eight‐node solid‐shell elements for geometric non‐linear analysis of elastic shells. To subdue shear, trapezoidal and thickness locking, the assumed natural strain method and an ad hoc modified generalized laminate stiffness matrix are employed. A selectively reduced integrated element is formulated with its membrane and bending shear strain components taken to be constant and equal to the ones evaluated at the element centroid. With the generalized stresses arising from the modified generalized laminate stiffness matrix assumed to be independent from the ones obtained from the displacement, an extended Hellinger–Reissner functional can be derived. By choosing the assumed generalized stresses similar to the assumed stresses of a previous solid element, a hybrid‐stress solid‐shell element is formulated. Commonly employed geometric non‐linear homogeneous and laminated shell problems are attempted and our results are close to those of other state‐of‐the‐art elements. Moreover, the hybrid‐stress element converges more readily than the selectively reduced integrated element in all benchmark problems. Copyright © 2002 John Wiley & Sons, Ltd.     

A 3D contact smoothing method using Gregory patches

In this work, a method is developed for smoothing three‐dimensional contact surfaces. The method can be applied to both regular and irregular meshes. The algorithm employs Gregory patches to interpolate finite element nodes and provide tangent plane continuity between adjacent patches. The resulting surface interpolation is used to calculate gaps and contact forces, in a variationally consistent way, such that contact forces due to normal and frictional contact vary smoothly as slave nodes transition from one patch to the next. This eliminates the ‘chatter’ which typically occurs in a standard contact algorithm when a slave node is situated near a master facet edge. The elimination of this chatter provides a significant improvement in convergence behaviour, which is illustrated by a number of numerical examples. Furthermore, smoothed surfaces also provide a more accurate representation of the actual surface, such that resulting stresses and forces can be more accurately computed with coarse meshes in many problems. This fact is also demonstrated by the numerical examples. Published in 2002 by John Wiley & Sons, Ltd.     

多约束的桥梁结构拓扑优化

基于ESO (Evolutionary Structural Optimization)的拓扑优化方法,本文提出了一种适应于桥梁结构的拓扑优化方法。引进了两种性能指标公式来确定最佳拓扑设计。另外,为了更有效地尝试改进最终设计的细节,而又不进行更精细有限元网格的完整分析,这里提出了一种精细网络设计方案。再者,结合细啃技术,完成了考虑应力、位移、频率约束的斜拉桥优化设计。通过几种桥梁的设计优化,表明该方法的可应用性、简洁性和有效性。