基于GA-WPA优化的BP神经网络目标威胁估计

在防空作战中,目标威胁估计是指挥控制过程的重要一环,是决策和指挥的重要依据。BP神经网络能够解决目标威胁估计问题,但存在收敛速度慢、易陷入局部最优等缺点。提出将遗传算法(Genetic Algorithm,GA)的选择、交叉和变异操作融入到狼群算法(Wolf Pack Algorithm,WPA)中,提出了GA-WPA算法,以提高狼群算法的收敛速度。在此基础上,利用所提出的GA-WPA算法对BP神经网络进行优化,确定最优初始权值和阈值。最后,将优化后的BP神经网络解决地面防空系统目标威胁估计问题。仿真实验表明,所提算法能够有效克服BP神经网络收敛速度慢、易陷入局部最优等缺点,能够提高目标威胁估计的准确性和适应性。     

Consistent empirical physical formula construction for recoil energy distribution in HPGe detectors by using artificial neural networks

The gamma-ray tracking technique is a highly efficient detection method in experimental nuclear structure physics. On the basis of this method, two gamma-ray tracking arrays, AGATA in Europe and GRETA in the USA, are currently being tested. The interactions of neutrons in these detectors lead to an unwanted background in the gamma-ray spectra. Thus, the interaction points of neutrons in these detectors have to be determined in the gamma-ray tracking process in order to improve photo-peak efficiencies and peak-to-total ratios of the gamma-ray peaks. In this paper, the recoil energy distributions of germanium nuclei due to inelastic scatterings of 1–5 MeV neutrons were first obtained by simulation experiments. Secondly, as a novel approach, for these highly nonlinear detector responses of recoiling germanium nuclei, consistent empirical physical formulas (EPFs) were constructed by appropriate feedforward neural networks (LFNNs). The LFNN-EPFs are of explicit mathematical functional form. Therefore, the LFNN-EPFs can be used to derive further physical functions which could be potentially relevant for the determination of neutron interactions in gamma-ray tracking process.     

Estimating the sound absorption coefficients of perforated wooden panels by using artificial neural networks

Developing efficient sound absorption materials is a relevant topic for large scale structures such as gymnasiums, shopping malls, airports and stations. This study employs artificial neural network (ANN) algorithm to estimate the sound absorption coefficients of different perforated wooden panels with various setting combinations including perforation percentage, backing material and thickness. The training data sets are built by carrying out a series of experimental measurements in the reverberation room to evaluate the sound absorption characteristics of perforated wooden panels. A multiple linear regression (MLR) model is also developed for making comparisons with ANN. The analytical results indicate that the ANN exhibits satisfactory reliability of a correlation between estimation and truly measured absorption coefficients of approximately 0.85. However, MLR cannot be applied to nonlinear cases. ANN is a useful and reliable tool for estimating sound absorption coefficients estimation.     

Integrating artificial neural networks and empirical correlations for the prediction of water-subcooled critical heat flux

The critical heat flux (CHF) is an important parameter for the design of nuclear reactors, heat exchangers and other boiling heat transfer units. Recently, the CHF in water-subcooled flow boiling at high mass flux and subcooling has been thoroughly studied in relation to the cooling of high-heat-flux components in thermonuclear fusion reactors. Due to the specific thermal-hydraulic situation, very few of the existing correlations, originally developed for operating conditions typical of pressurized water reactors, are able to provide consistent predictions of water-subcooled-flow-boiling CHF at high heat fluxes. Therefore, alternative predicting techniques are being investigated. Among these, artificial neural networks (ANN) have the advantage of not requiring a formal model structure to fit the experimental data; however, their main drawbacks are the loss of model transparency (‘black-box’ character) and the lack of any indicator for evaluating the accuracy and reliability of the ANN answer when ‘never-seen’ patterns are presented. In the present work, the prediction of CHF is approached by a hybrid system which couples a heuristic correlation with a neural network. The ANN role is to predict a datum-dependent parameter required by the analytical correlation; this parameter was instead set to a constant value obtained by usual best-fitting techniques when a pure analytical approach was adopted. Upper and lower boundaries can be possibly assigned to the parameter value, thus avoiding the case of unexpected and unpredictable answer failure. The present approach maintains the advantage of the analytical model analysis, and it partially overcomes the ‘black-box’ character typical of the straight application of ANNs because the neural network role is limited to the correlation tuning. The proposed methodology allows us to achieve accurate results and it is likely to be suitable for thermal-hydraulic and heat transfer data processing.     

Application of artificial neural networks for unfolding neutron spectra by using a scintillation detector

The unfolding of neutron spectra from the pulse height distribution measured by a BC501A scintillation detector is accomplished by the application of artificial neural networks (ANN). A simple linear neural network without biases and hidden layers is adopted. A set of monoenergetic detector response functions in the energy range from 0.25 MeV to 16 MeV with an energy interval of 0.25 MeV are generated by the Monte Carlo code O5S in the training phase of the unfolding process. The capability of ANN was demon...     

Influence of holmium doping on the optical properties of quaternary InGaAsP epitaxial layers

Photoluminescence and photoconductivity measurements were used to study the influence of Ho doping on the optical properties of InGaAsP layers grown by liquid phase epitaxy (LPE). The full width at half maximum (FWHM) of the photoluminescence peak was found to decrease as the amount of Ho increases. When the amount of Ho is 0.11 wt%, the FWHM has a minimum value of 7.93 meV, about 46% lower than that of the undoped InGaAsP. The absorption tails observed in the photoconductivity were analyzed with the Urbach tail model and the Urbach energies were obtained from the fits. The Urbach energy decreases as the amount of Ho increases, indicating that Ho doping greatly reduces the amount of residual impurities in LPE-grown layers.     

Prediction of concrete strength using ultrasonic pulse velocity and artificial neural networks

Ultrasonic pulse velocity technique is one of the most popular non-destructive techniques used in the assessment of concrete properties. However, it is very difficult to accurately evaluate the concrete compressive strength with this method since the ultrasonic pulse velocity values are affected by a number of factors, which do not necessarily influence the concrete compressive strength in the same way or to the same extent. This paper deals with the analysis of such factors on the velocity-strength relationship. The relationship between ultrasonic pulse velocity, static and dynamic Young’s modulus and shear modulus was also analyzed. The influence of aggregate, initial concrete temperature, type of cement, environmental temperature, and w/c ratio was determined by our own experiments. Based on the experimental results, a numerical model was established within the Matlab programming environment. The multi-layer feed-forward neural network was used for this purpose. The paper demonstrates that artificial neural networks can be successfully used in modelling the velocity-strength relationship. This model enables us to easily and reliably estimate the compressive strength of concrete by using only the ultrasonic pulse velocity value and some mix parameters of concrete.     

Modelling the soundscape quality of urban waterfronts by artificial neural networks

The renewal of the urban waterfronts has become a major focus of attention for politicians and decision makers in the city’s management programs. The recognition of the patterns that define the waterfronts’ identity is essential to select new strategies of intervention for the environmental recovery. In order to create adequate environments for everyday life within a sustainable development, new links between human senses, human perception and design need to be created. Within this wide approach, the landscape and the soundscape play a significant role and can become a key driving force in the implementation of the changes. New techniques have to be tested to identify the sonic and visual parameters capable to explain the specificity of a waterfront. With this purpose, an artificial neural network (ANN) was developed, and the relative importance of the input variables was evaluated. The collected database was also analysed by multiple linear regression (MLR) to compare the outcomes of both models. The urban waterfront of Naples (Italy) was chosen as case study. The results obtained show that the performance of the neural network is better than the one of the linear regression (rANN = 0.949, rMLR = 0.639). The interpretation of the relative importance method is also quite satisfactory in the ANN.     

变参数混沌时间序列的神经网络预测研究

研究一类复杂变参数混沌系统时间序列的预测问题.首先构造一个变参数Logistic映射,分析变参数混沌系统的特点,指出动力学特征不断变化的这类系统不存在恒定形状的吸引子;结合Takens嵌入定理和神经网络理论,阐述神经网络方法预测具有恒定吸引子形状的混沌系统可行的原因,分析研究其用于预测变参数混沌系统的潜在问题.变参数Ikeda系统的神经网络预测试验验证了理论分析结果,试验还表明,简单增大预测训练样本数可能降低泛化预测精度,训练集的选择对这类系统的泛化预测效果影响极大,指出混沌时间序列预测实用化必须研究解决这类变参数混沌系统的预测. 关键词： 混沌 预测 神经网络 变参数系统     

Theoretical prediction of structural, electronic and optical properties of quaternary alloy Zn1-xBexSySe1-y

Within density functional theory based on the full potential-linearized augmented plane wave method,we carry out the first-principles calculation of the structural,electronic,and optical properties of the zinc blende quaternary alloy Zn_1-xBe_xS_ySe_1-y.The Perdew-Burke-Ernzerhof generalized gradient approximation based on the optimization of total energy and the Engel-Vosko generalized gradient approximation based on the optimization of the corresponding potential are used.Our investigation on the effect of the composition on lattice constants,bulk modulus,band gap,optical dielectric constant,and refractive index shows a non-linear dependence.The energy gap E g（x,y） has been determined over the entire compositions x and y.In addition,the energy band gap of the technologically important quaternary alloy Zn_1-xBe_xS_ySe_1-y in conditions of being lattice matched to ZnS substrate has been investigated.It is noteworthy that the present work is the first theoretical study of the quaternary alloy of interest.     

Determination of some rare earth elements by EDXRF and artificial neural networks

This paper describes the simultaneous determination of Pr, Nd and Sm by EDXRF spectrometry using mixtures of oxides of these metals in a silica matrix. The data were treated by distinct neural network algorithms: back‐propagation (BP), Levenberg–Marquardt (LM) and two variations of back‐propagation (called BP‐SC, single component, and BP‐MC, multiple component), using results from the PLS model (partial least square regression) for comparison. The best applied model was the BP‐SC neural network, which produced relative standard errors of prediction of 17.5% for Pr, 12.5% for Nd and 12.6% for Sm. Copyright © 2003 John Wiley & Sons, Ltd.     

Estimation and interpolation of underwater low frequency ambient noise spectrum using artificial neural networks

In this work, estimation of ambient noise spectrum influenced by wind speed and wave height carried out for the frequency range of 500 Hz to 5 kHz using Feed forward Neural Network (FNN) is presented. Ocean ambient noise measurements were made in the shallow waters of Bay of Bengal using a portable data acquisition system with a high sensitivity hydrophone at a depth of 5 m from the surface.100 sets of data covering a rage of wind speeds from 2.5 m/s to 8.5 m/s with approximately 15 sets of data falling within 1 m/s over the range of wind speed were used for training the FNN. The parameter wave height which contributes to the noise producing mechanism is also used for training along with wind speed. The results revealed that the proposed method is useful in the estimation and interpolation of underwater noise spectrum level and hence in simulation for the considered frequency range. These were confirmed by calculating the Mean Squared Error (MSE) between the experimental data and the simulation. As the measurements of the underwater ambient noise level are very difficult in remote oceanic regions, where conditions are often inhospitable, these studies seem to be relevant.     

Nonlinear dynamics extraction for time-delay systems using modular neural networks synchronization and prediction

It is shown that the nonlinear dynamics of chaotic time-delay systems can be reconstructed using a new type of neural network with two modules: one for nonfeedback part with input data delayed by the embedding time, and a second one for the feedback part with input data delayed by the feedback time. The method is applied to both simulated and experimental data from an electronic analog circuit of the Mackey–Glass system. Better results are obtained for the modular than for feedforward neural networks for the same number of parameters. It is found that the complexity of the neural network model required to reconstruct nonlinear dynamics does not increase with the delay time. Synchronization between the data and the model with diffusive coupling is also achieved. We have also shown by iterating the model from the present point that the dynamics can be predicted with a forecast horizon larger than the feedback delay time.     

Optimal design for laser beam butt welding process parameter using artificial neural networks and genetic algorithm for super austenitic stainless steel

Laser welding input parameters play a very significant role in determining the quality of a weld joint. The joint quality can be defined in terms of properties such as weld bead geometry, mechanical properties and distortion. Therefore, mechanical properties should be controlled to obtain good welded joints. In this study, the weld bead geometry such as depth of penetration (DP), bead width (BW) and tensile strength (TS) of the laser welded butt joints made of AISI 904L super austenitic stainless steel were investigated. Full factorial design was used to carry out the experimental design. Artificial Neural networks (ANN) program was developed in MatLab software to establish the relationships between the laser welding input parameters like beam power, travel speed and focal position and the three responses DP, BW and TS in three different shielding gases (Argon, Helium and Nitrogen). The established models were used for optimizing the process parameters using Genetic Algorithm (GA). Optimum solutions for the three different gases and their respective responses were obtained. Confirmation experiment has also been conducted to validate the optimized parameters obtained from GA.     

The electrical impedance,AC conductivity and dielectric properties of phenol red compound investigated and modeled by an artificial neural network

The impedance spectroscopy, electrical conductivity and electric modulus of bulk phenol red were measured, as a function of both frequency and temperature. Artificial neural networks (ANNs) were used for modeling its electrical properties. The two parts (real and imaginary) of its complex impedance (Z*) were analyzed and the activation energy related to the electrical relaxation process was evaluated. Nyquist curves were plotted showing semicircles for the different temperatures. The AC electrical conductivity follows a power law σ_ac(ω) α ω^η. The maximum barrier height B_m was derived for specific temperatures. A plausible mechanism for the AC conduction of bulk phenol red was deduced from the temperature reliance of the frequency exponent. The dielectric data was analyzed using electric modulus as a tool. In addition, ANNs were used to model the impedance parts and the total electrical conductivity. Numerous runs were tried, to obtain the best performance. The training and prediction results were compared to the equivalent experimental results, with a good match obtained. An equation describing the experimental results was obtained mathematically, based on the use of ANNs. The outputs demonstrated that ANNs are an admirable tool for modeling experimental results.     

Wooden windows: Sound insulation evaluation by means of artificial neural networks

Windows are the weakest part of a façade in terms of acoustic performance: the weighted sound insulation index (R_w), measured according to ISO 140-3, is the fundamental parameter to evaluate the façade acoustic insulation.The paper aims at developing an artificial neural network (ANN) model to estimate the R_w value of wooden windows based on a limited number of windows parameters; this is a new approach because acoustic phenomena are non-linear and affected by a plurality of factors and, therefore, usually investigated through experimentation.Data set is taken from experimental campaigns carried out at the Laboratory of Acoustics, University of Perugia. A multilayer feed-forward approach was chosen and the model was implemented in MATLAB. On the basis of the results obtained by means of a preliminary training and test campaign of several ANN architectures, five main parameters were selected as network inputs: window typology, frame and shutters thickness, number of gaskets, R_w of glazing; R_w value of the window is the network output. Different ANN configurations were trained and a root mean-square error less than 3% was obtained, comparable to measurement uncertainty.This approach allows to develop a model which, with input parameters varying within appropriate ranges, can easily estimate the acoustic performance of wooden windows without experimental campaign on prototypes, saving both money and time. If the training data set is large enough, the presented approach could be very useful for design and optimization of acoustic performance of new products.     

The investigation of model selection criteria in artificial neural networks by the Taguchi method

Artificial neural networks (ANNs) have been successfully used for solving variety of problems. One major disadvantage of ANNs is that there is no formal systematic model building approach. This paper presents the application of the Taguchi method in the optimization of the design parameters of the ANNs. The performances of the ANNs were determined by the Taguchi method considering factors relevant for ANNs’ performance. The properties affecting the performance of the ANNs and their levels on the peak analytical function were determined by performing computational experiments. After training the network, the values of the statistical data criteria were determined and the optimum parameter levels were obtained in terms of the performance statistics. The performance of ANNs is shown to be better in the case of the application of the Taguchi method rather than in the case of random choice of factor values.     

人工神经网络在建筑声学中的应用

本文对人工神经网络的历史作了简单回顾，并对它在建筑声学领域中的应用状况作了阐述，同时提出了有待进一步探讨的若干问题。     

Detection of material property errors in handbooks and databases using artificial neural networks with hidden correlations

The authors have discovered a systematic, intelligent and potentially automatic method to detect errors in handbooks and stop their transmission using unrecognised relationships between materials properties. The scientific community relies on the veracity of scientific data in handbooks and databases, some of which have a long pedigree covering several decades. Although various outlier-detection procedures are employed to detect and, where appropriate, remove contaminated data, errors, which had not been discovered by established methods, were easily detected by our artificial neural network in tables of properties of the elements. We started using neural networks to discover unrecognised relationships between materials properties and quickly found that they were very good at finding inconsistencies in groups of data. They reveal variations from 10 to 900% in tables of property data for the elements and point out those that are most probably correct. Compared with the statistical method adopted by Ashby and co-workers [Proc. R. Soc. Lond. Ser. A 454 (1998) p. 1301, 1323], this method locates more inconsistencies and could be embedded in database software for automatic self-checking. We anticipate that our suggestion will be a starting point to deal with this basic problem that affects researchers in every field. The authors believe it may eventually moderate the current expectation that data field error rates will persist at between 1 and 5%.     

Improving excitation and inversion accuracy by optimized RF pulse using genetic algorithm

In this study, a Genetic Algorithm (GA) is introduced to optimize the multidimensional spatial selective RF pulse to reduce the passband and stopband errors of excitation profile while limiting the transition width. This method is also used to diminish the nonlinearity effect of the Bloch equation for large tip angle excitation pulse design. The RF pulse is first designed by the k-space method and then coded into float strings to form an initial population. GA operators are then applied to this population to perform evolution, which is an optimization process. In this process, an evaluation function defined as the sum of the reciprocal of passband and stopband errors is used to assess the fitness value of each individual, so as to find the best individual in current generation. It is possible to optimize the RF pulse after a number of iterations. Simulation results of the Bloch equation show that in a 90 degrees excitation pulse design, compared with the k-space method, a GA-optimized RF pulse can reduce the passband and stopband error by 12% and 3%, respectively, while maintaining the transition width within 2 cm (about 12% of the whole 32 cm FOV). In a 180 degrees inversion pulse design, the passband error can be reduced by 43%, while the transition is also kept at 2 cm in a whole 32 cm FOV.