RBMDO Using Gaussian Mixture Model-Based Second-Order Mean-Value Saddlepoint Approximation

Actual engineering systems will be inevitably affected by uncertain factors.Thus,the Reliability-Based Multidisciplinary Design Optimization(RBMDO)has become a hotspot for recent research and application in complex engineering system design.The Second-Order/First-Order Mean-Value Saddlepoint Approximate(SOMVSA/-FOMVSA)are two popular reliability analysis strategies that are widely used in RBMDO.However,the SOMVSA method can only be used efficiently when the distribution of input variables is Gaussian distribution,which significantly limits its application.In this study,the Gaussian Mixture Model-based Second-Order Mean-Value Saddlepoint Approximation(GMM-SOMVSA)is introduced to tackle above problem.It is integrated with the Collaborative Optimization(CO)method to solve RBMDO problems.Furthermore,the formula and procedure of RBMDO using GMM-SOMVSA-Based CO(GMM-SOMVSA-CO)are proposed.Finally,an engineering example is given to show the application of the GMM-SOMVSA-CO method.

A fast space-time-domain Gaussian beam migration approach using the dominant frequency approximation

The Gaussian beam migration(GBM) is a steady imaging approach, which has high accuracy and efficiency. Its implementation mainly includes the traditional frequency domain and the recent popular space-time domain. Firstly, we use the upward ray tracing strategy to get the backward wavefields. Then,we use the dominant frequency of the seismic data to simplify the imaginary traveltime calculation of the wavefields, which can cut down the Fourier transform number compared with the traditional GBM in the space-time domain. In addition, we choose an optimized parameter for the take-off angle increment of the up-going and down-going rays. These optimizations help us get an efficient space-time-domain acoustic GBM approach. Typical four examples show that the proposed method can significantly improve the computational efficiency up to one or even two orders of magnitude in different models with different model parameters and produce good imaging results with comparable accuracy and resolution with the traditional GBM in the space-time domain.

Mode-field half-widths of Gaussian approximation for the fundamental mode of two kinds of optical waveguides

This paper analyzes the characteristic of matching efficiency between the fundamental mode of two kinds of optical waveguides and its Gaussian approximate field.Then, it presents a new method where the mode-field half-width of Caussian approximation for the fundamental mode should be defined according to the maximal matching efficiency method. The relationship between the mode-field half-width of the Gaussian approximate field obtained from the maximal matching efficiency and normalized frequency is studied; furthermore, two formulas of mode-field half-widths as a function of normalized frequency are proposed.

Theoretical analysis of pilotless frame synchronization for LDPC code using Gaussian approximation

The pilotless frame synchronization approach and implementations of LDPC code are the crucial issue of LDPC decoder. The Maximum-A-Posteriori probability( MAP) decoder has a perfect frame synchronization error rate( FSER) performance. In this paper,a theoretical derivation of the FSER performance of pilotless frame synchronization for LDPC code is presented. The FSER performance by theoretical analysis coincides well with that by simulation in additive white Gaussian channel and Rician fading channel. So it is estimated the FSER performance of an LDPC code by theoretical analysis can be used instead of the simulations which are much more time-consuming.

Analysis on Far-Field of Slab Waveguide and Its Gaussian Approximation

This paper researches end diffraction of slab waveguide and then matching efficiency between the far-field and its Gaussian approximate field is analyzed leads to a new definition of divergence half-angle. Finally, why the far-field can be approximated by a Gaussian function is presented according to characteristic of beam propagation factor.

An Approximate Analytical Propagation Formula for Gaussian Beams through a Cat-Eye Optical Lens under Large Incidence Angle Condition

Based on the generalized diffraction integral formula and the idea that the angle misalignment of the cat-eye optical lens can be transformed into the displacement misalignment,an approximate analytical propagation formula for Gaussian beams through a cat-eye optical lens under large incidence angle condition is derived.Numerical results show that the diffraction effect of the apertures of the cat-eye optical lens becomes stronger along with the increase in incidence angle.The results are also compared with those from using an angular spectrum diffraction integral and experiment to illustrate the applicability and validity of our theoretical formula.It is shown that the approximate extent is good enough for the application of a cat-eye optical lens with a radius of 20 mm and a propagation distance of 100 m,and the approximate extent becomes better along with the increase in the radius of the cat-eye optical lens and the propagation distance.

THE REGULARIZED SOLUTION APPROXIMATION OF FORWARD/BACKWARD PROBLEMS FOR A FRACTIONAL PSEUDO-PARABOLIC EQUATION WITH RANDOM NOISE

This paper deals with the forward and backward problems for the nonlinear fractional pseudo-parabolic equation ut+(-Δ)^(s1)ut+β(-Δ)^(s2)u=F(u,x,t)subject o random Gaussian white noise for initial and final data.Under the suitable assumptions s1,s2andβ,we first show the ill-posedness of mild solutions for forward and backward problems in the sense of Hadamard,which are mainly driven by random noise.Moreover,we propose the Fourier truncation method for stabilizing the above ill-posed problems.We derive an error estimate between the exact solution and its regularized solution in an E‖·‖Hs22norm,and give some numerical examples illustrating the effect of above method.

Robust Frequency Estimation Under Additive Symmetric α-Stable Gaussian Mixture Noise

Here the estimating problem of a single sinusoidal signal in the additive symmetricα-stable Gaussian(ASαSG)noise is investigated.The ASαSG noise here is expressed as the additive of a Gaussian noise and a symmetricα-stable distributed variable.As the probability density function(PDF)of the ASαSG is complicated,traditional estimators cannot provide optimum estimates.Based on the Metropolis-Hastings(M-H)sampling scheme,a robust frequency estimator is proposed for ASαSG noise.Moreover,to accelerate the convergence rate of the developed algorithm,a new criterion of reconstructing the proposal covar-iance is derived,whose main idea is updating the proposal variance using several previous samples drawn in each iteration.The approximation PDF of the ASαSG noise,which is referred to the weighted sum of a Voigt function and a Gaussian PDF,is also employed to reduce the computational complexity.The computer simulations show that the performance of our method is better than the maximum likelihood and the lp-norm estimators.

基于GA-MP的低复杂度OTFS检测算法

针对正交时频空(Orthogonal Time Frequency Space,OTFS)通信系统信号检测复杂度高的问题,提出一种改进的高斯近似消息传递(Gaussian Approximate Message Passing,GA-MP)检测算法。依据最大后验概率检测准则,对发送信号及隐变量进行逐符号高斯近似,基于置信传播算法与联合因子图进行消息传递,用边缘后验概率替代GA-MP中的外部信息以减少运算量,结合阻尼因子提升收敛速度,同时引入概率阈值减少后续更新的节点数,从而使运算复杂度得到有效降低。实验结果表明,改进后的GA-MP算法在保证误码率性能的前提下具有更低的复杂度。

基于GMM和GA-LSTM的稀土熔盐电解过程原料含量状态识别模型

在高温高风险的稀土熔盐电解工艺中,为了实现稀土熔盐电解过程原料含量状态的智能识别,提出了一种基于混合高斯背景建模(GMM)和遗传算法优化的长短期记忆神经网络(GA-LSTM)的分类模型。模型通过GMM算法、R通道自适应滤波和中值滤波准确提取图像的火焰前景和特征,以量化熔盐电解反应的剧烈程度,进而判断稀土熔盐电解处于原料含量过多或含量正常状态;然后利用GA-LSTM神经网络建立熔盐表面火焰特征和稀土熔盐电解过程原料含量状态的非线性映射关系。结果表明:模型的识别精度高达99.79%,具有较好的泛化性,为实现稀土熔盐电解工艺自动化提供了一定的参考价值。

Cell Gas Free Energy as an Approximation of the Continuous Model

A continuous infinite system of point particles interacting via two-body strong superstable potential is considered in the framework of cell gas (CG) model of classical statistical mechanics. We consider free energy of this model as an approximation of the correspondent value of the continuous system. It converges to the free energy of the conventional continuous gas if the parameter of approximation α→0 for any values of an inverse temperature β>0 and volume per particle ν>0.

利用Gaussian核对多元函数的近似逼近及其误差估计

V.Maz’ya首次提出了近似逼近法,其主要是研究定义在全空间上的光滑函数的逼近情况,但它不能有效的处理积分和拟微分算子的高阶求积公式问题及利用更有效的数值和半数值方法解决数学物理的边界等问题.F.M櫣ller和W.Varnhorn给出了一维紧区间上函数的近似逼近方法,而且还可以控制近似逼近的截断误差.根据上述思想,采用近似逼近法,利用Gaussian核对二维紧空间上光滑函数进行逼近,并考察由这种近似逼近法所产生的误差情况.

基于Gaussian型RBF神经网络的一元函数逼近性能研究

为了研究Gaussian型RBF神经网络对于一元非线性函数的逼近能力,编程建立了Gaussian型RBF神经网络和BP神经网络,并以正弦函数、指数函数、阶跃函数三种典型的一元非线性函数为例,分别用两种神经网络对其进行逼近.仿真结果表明,相对于传统BP神经网络而言,Gaussian型RBF神经网络对于一元非线性函数的逼近精度更高、收敛速度更快,具有良好的逼近能力,为解决一元非线性函数的逼近问题提供了良好的解决手段.

利用Gaussian型RBF网络进行函数逼近的构造性估计

前馈人工神经网络有着极其广泛的应用 ,如何估计隐层神经元数及相应的逼近误差 ,一直是确定前馈网络结构的难点和关键。 RBF网络是一种最重要的前馈神经网络 ,本文给出了利用 Gaussian型 RBF网络逼近连续函数或 Lebesgue-可积函数时的构造性的隐层单元数显式估算式及相应的显式逼近误差估算式。文中的结论也易于推广到离散样本的情形。这些结论对于提高 Guassian型 RBF在实际应用时的计算精度和减少计算量具有一定的指导意义。

Gaussian型RBF神经网络的函数逼近仿真研究

为了研究Gaussian型RBF神经网络的逼近能力,首先介绍了Gaussian型RBF神经网络的结构和算法,然后在MATLAB7.0环境下,编程建立了Gaussian型RBF神经网络和BP神经网络,并以具体的非线性函数为例,分别用两种神经网络对其进行逼近.仿真结果表明,相对于传统BP神经网络而言,Gaussian型RBF神经网络对于非线性函数的逼近精度更高、收敛速度更快,具有良好的逼近能力,为解决非线性函数的逼近问题提供了良好的解决手段.

基于Plane-Gaussian神经网络的网络流状态监测

针对复杂网络环境下网络流监测(分类)问题,为实现多个类别直接分类以及提高学习方法的训练速度,提出了一种随机的人工神经网络学习方法。该方法借鉴平面高斯(PG)神经网络模型,引入随机投影思想,通过计算矩阵伪逆的方法解析获得网络连接矩阵,理论上可证明该网络具有全局逼近能力。在人工数据和标准网络流监测数据上进行了实验仿真,与同样采用随机方法的极限学习机(ELM)和PG网络相比,分析与实验结果表明:1)由于继承了PG网络的几何特性,对平面型分布数据更为有效;2)采用了随机方法,训练速度与ELM相当,但比PG网络快得多;3)三种方法中,该方法更有利于解决网络流监测问题。

基于高斯近似和NSGA-Ⅱ的伴流螺旋桨优化

[目的]为改进船舶推进性能,提高船舶效能,克服螺旋桨参数众多、建模繁琐、水动力数值计算收敛速度慢、伴流场的湍流效应等问题,提出一套螺旋桨高效优化方法。[方法]首先,将基于非均匀有理B样条的自由曲面变形(NFFD)技术构建的参数化螺旋桨作为输出,利用CFD数值仿真对螺旋桨性能特征进行预报。然后,基于仿真数据建立螺旋桨性能高斯近似预测模型,建立以提高效率、降低扭矩系数为目标的优化模型,利用二代非支配排序遗传算法(NSGA-Ⅱ)得到最优解。[结果]结果显示,在均匀流场和船后伴流场下,螺旋桨的优化结果有着相似规律,即桨叶宽度和厚度均呈减小的趋势,倾斜角呈增大趋势,但由于在船后伴流场下流速会降低,因而桨叶宽度和厚度的减小以及倾斜角的增加幅度相对于均匀流场下的小。[结论]采用所构建的方法可以实现均匀流场和船后伴流场下螺旋桨的高效优化设计。

基于GPR代理模型和GA-APSO混合优化算法的软基水闸底板脱空反演

软基水闸底板脱空是水闸在长期服役期间受水流侵蚀等环境因素影响所产生的一种危害极大且难以察觉的病害。由于其病害部位于水下,传统方法难以检测,该研究提出一种基于高斯过程回归(Gaussian process regression,GPR)代理模型和遗传-自适应惯性权重粒子群(genetic algorithm-adaptive particle swarm optimization,GA-APSO)混合优化算法的水闸底板脱空动力学反演方法,用于检测软基水闸底板脱空。首先,构建表征软基水闸底板脱空参数和水闸结构模态参数之间非线性关系的GPR代理模型;其次,基于GPR代理模型与水闸实测模态参数建立脱空反演的最优化数学模型,将反演问题转化为目标函数最优化求解问题;最后,为提高算法寻优计算的精度,提出一种GA-APSO混合优化算法对目标函数进行脱空反演计算,并提出一种更合理判断反演脱空区域面积和实际脱空区域面积相对误差的指标—面积不重合度。为验证所提方法性能,以一室内软基水闸物理模型为例,对两种不同脱空工况开展研究分析,结果表明,反演脱空区域面积和模型实际设置脱空区域面积的相对误差分别为8.47%和10.77%,相对误差值较小,证明所提方法能有效反演出水闸底板脱空情况,可成为软基水闸底板脱空反演检测的一种新方法。

ON THE NECESSARY AND SUFFICIENT CONDITIONS TO SOLVE A HEAT EQUATION WITH GENERAL ADDITIVE GAUSSIAN NOISE

In this note, we consider stochastic heat equation with general additive Gaussian noise. Our aim is to derive some necessary and sufficient conditions on the Gaussian noise in order to solve the corresponding heat equation. We investigate this problem invoking two differen t met hods, respectively, based on variance compu tations and on pat h-wise considerations in Besov spaces. We are going to see that, as anticipated, both approaches lead to the same necessary and sufficient condition on the noise. In addition, the path-wise approach brings out regularity results for the solution.

Electromagnetic scattering from two-layer rough interfaces in the Kirchhoff approximation

Electromagnetic wave scattering from multilayers consisting of two two-layer Caussian rough surfaces with lossless media is investigated in the Kirchhoff approximation （KA）, with consideration of the shadowing effects. The tapered incident wave is introduced into the classic KA, and the bistatic scattering coefficient is redetermined. The advantage of this method is that it is faster in computation than the exact numerical methods. The numerical results show that the bistatic scattering coefficient calculated in the KA is in good agreement with that obtained by using the method of moment （MOM） over a most angular range, which indicates the validity of the KA proposed in our paper. Finally, the effects of the relative permittivity, the root-mean-square （RMS） height, the correlative length, and the average height between the two interfaces on the bistatic scattering coefficient are discussed in detail.