基于GASA混合策略的BP网络在基准地价测算中的应用

结合遗传算法的并行搜索结构和模拟退火的概率突跳特性 ,提出了一种用于BP网络权值学习的GASA混合策略。以江西省赣县县城商业用地为例 ,应用基于GASA混合策略的BP网络对其基准地价进行了测算 ,并与回归模型方法作了比较。结果表明 ,混合策略能有效地避免BP算法陷入局部极小和网络单目标学习易产生的过拟合现象。将神经网络用于基准地价的测算 ,精度优于常规的回归模型方法。     

Optimum design of structures against earthquake by wavelet neural network and filter banks

Optimum design of structures for earthquake is achieved by simulated annealing. To reduce the computational work, a fast wavelet transform is used by means of which the number of points in the earthquake record is decreased. The record is decomposed into two parts. One part contains the low frequency of the record, and the other contains the high frequency of the record. The low‐frequency content is the effective part, since most of the energy of the record is contained in this part of the record. Thus, the low‐frequency part of the record is used for dynamic analysis. Then, using a wavelet neural network, the dynamic responses of the structures are approximated. By such approximation, the dynamic analysis of the structure becomes unnecessary in the process of optimization. The wavelet neural networks have been employed as a general approximation tool for the time history dynamic analysis. A number of structures are designed for optimal weight and the results are compared to those corresponding to the exact dynamic analysis. Copyright © 2004 John Wiley & Sons, Ltd.     

确定河流纵向离散系数的快速SA法

将模拟退火算法应用求解分析河流水团示踪试验数据,确定河流纵向离散系数的函数优化问题.针对一般模拟退火算法收敛性能差的缺陷,采取了改变模型扰动、接受概率、降温方式以及增加附加约束条件和增加记忆功能等综合措施对算法进行了改进.实例分析结果表明,改进后的模拟退火算法确定河流纵向离散系数的优化问题不仅可行,而且求解精度较高,更重要的是收敛速度比其它算法快的多,这在一些高维复杂的优化问题中很值得推广.     

裂变径迹定年方法的进展及应用

裂变径迹分析方法利用裂变径迹年龄及裂变径迹长度分布形态来反映随时间产生的裂变径迹积累及由温度控制的退火(裂变径迹缩短或消失)两种因素综合作用的结果。近几年来，有关磷灰石的退火特征进行了大量实验和野外研究，取得了较大进展。本文通过对80年代以来国内外裂变径迹研究状况的追踪调研结果，简要介绍裂变径迹定年方法的主要进展及应用情况。     

Source tomography by simulated annealing using broad-band surface waves and geodetic data: application to the Mw= 8.1 Chile 1995 event

We invert surface-wave and geodetic data for the spatio-temporal complexity of slip during the M _w =8.1 Chile 1995 event by simulated annealing. This quasi-global inversion method allows for a wide exploration of model space, and retains the non-linearity of the source tomography problem. Complex source spectra are obtained from 5 to 45 mHz from first- and second-orbit fundamental-mode Rayleigh waves using an empirical Green's function cross-correlation technique. Coseismic displacement vectors were measured at 10 GPS sites near Antofagasta. They are part of a French-Chilean experiment which monitors the Northern Chile seismic gap. The spectra, together with the geodetic data, are inverted for the moment distribution on a 2-D dipping fault, under the physical constraints of slip positivity and causality. Marginal a posteriori distributions of the model parameters are obtained from several independently inverted solutions. In general, features of the slip model are well resolved. Data are well fitted by a purely unilateral southward rupture with a nearly uniform velocity around 2.5–3.0 km s⁻¹, and a total duration of 65 s. Several regions of moment release were imaged, one near the hypocentre, a major one 80 km south of it and a minor one 160 km south of it. The major patch of moment release seemed to have propagated to relatively shallow depths near the trench, 100 km SSW of the epicentre. The region of major slip is located updip of the 1987, M _w =7.5 earthquake, suggesting a causal relationship. Most of the slip occurred updip of the hypocentre (36 km), but the entire coupled plate interface (20–40 km) ruptured during the Chile 1995 event.     

ROBUST PRINCIPAL COMPONENT ANALYSIS BY PROJECTION PURSUIT

Principal component analysis (PCA) is a widely used technique in chemometrics.The classical PCAmethod is,unfortunately,non-robust,since the variance is adopted as the objective function.In thispaper,projection pursuit (PP) is used to carry out PCA with a criterion which is more robust than thevariance.In addition,the generalized simulated annealing (GSA) algorithm is introduced as anoptimization procedure in the process of PP calculation to guarantee the global optimum.The resultsfor simulated data sets show that PCA via PP is resistant to the deviation of the error distribution fromthe normal one.The method is especially recommended for use in cases with possible outlier(s) existingin the data.     

Joint inversion of travel times and waveforms by means of annealing algorithms. Application to a seismic refraction experiment

An algorithm of annealing is applied to a joint inversion of travel times and waveforms belonging to a synthetically generated seismic refraction experiment. The medium (crust and upper mantle) is modelled by a set of plane stratified layers and a halfspace. The obtained structure (elastic parameters and depth of layers) shows that, in spite of contaminating seismic noise and poor knowledge of the seismic source, annealing methods are a good tool in these kinds of inversion problems. We think that many characteristics of the annealing process described here could be used with real data and more sophisticated media for the crust and upper mantle of the earth than the present example.     

Practical considerations in the application of simulated annealing to stochastic simulation

Realizations generated by conditional simulation techniques must honor as much data as possible to be reliable numerical models of the attribute under study. The application of optimization methods such as simulated annealing to stochastic simulation has the potential to honor more data than conventional geostatistical simulation techniques. The essential feature of this approach is the formulation of stochastic imaging as an optimization problem with some specified objective function. The data to be honored by the stochastic images are coded as components in a global objective function. This paper describes the basic algorithm and then addresses a number of practical questions: (1) what are the criteria for adding a component to the global objective function? (2) what perturbation mechanism should be employed in the annealing simulation? (3) when should the temperature be lowered in the annealing procedure? (4) how are edge/border nodes handled? (5) how are local conditioning data handled? and (6) how are multiple components weighted in the global objective function?     

混沌退火Hopfield网络及其在储层聚类分析中的应用

将一种基于混沌退火的Hopfield神经网络(ACHN)应用到储层聚类分析中。这里主要阐述了ACHN聚类的基本原理、混沌神经元的标定、ACHN聚类的策略、先验信息的约束以及混沌退火搜索。实际砂砾岩储层聚类分析表明:这种网络能克服普通Hopfield网络陷入局部最优的不足,得到更准确、精细的聚类结果。