Bayes估计的进一步研究及其Pitman最优性

对线性模型参数,讨论了Bayes估计的Pitman最优性,将已有结果进行了改进,去掉了附加条件,证明了在Pitman准则下,Bayes估计一致优于最小二乘估计(LSE),在此基础上,提出了一种基于先验信息的方差分量估计,通过和基于LSE的方差分量估计作比较,证明了新估计是无偏估计且有更小的均方误差．最后,证明了在Pitman准则下生长曲线模型参数的Bayes估计优于最佳线性无偏估计．     

4种等温物态方程适用性的比较研究

 通过分析现有文献，认为4种两参数的物态方程在高压物理研究中较为流行，它们是Vinet、Baonza、Morse和Born-Meyer（BM）方程。将这4种方程用于拟合50种材料的实验压缩数据，得出了零压下压缩模量及其一阶压力导数的优化取值，并计算了4种方程的平均压力误差。结果表明，Morse方程的精度最高，对50种材料的平均拟合误差为0.557 6%；BM方程次之，平均误差为0.615 1%；Vinet和Baonza方程的误差大一些，分别为0.788 9%和0.833 3%。对8种典型材料计算了压力误差随压强变化的曲线，所得结果与平均误差的趋势一致，也是Morse方程的精度最高。在宽广压力范围的高压物态方程研究中，推荐使用Morse方程。     

Isomorphisms of fiber spaces over Teichmüller spaces

We will be mainly concerned with some important fiber spaces over Teichmuller spaces, including the Bers fiber space and Teichmuller curve, establishing an isomorphism theorem between "punctured" Teichmuller curves and determining the biholomorphic isomorphisms of these fiber spaces.     

Numerical analysis for the periodic response of nonsmooth dynamic systems of high-dimension

This paper deals with the numerical solution of high dimensional dynamic systems with nonsmooth characteristics, such as structural elastoplasticity, mechanical clearance and dry friction. As a stable periodic response of the system attracts its transient response, it may be extracted from the transient response somehow. The paper suggests a novel iteration scheme to extrapolate the periodic response from a short time history of its transient response by curve fitting. Compared with the current schemes such as shooting and incremental harmonic balancing, the present scheme makes full use of the information within the transient response and the system characters, then its computation efficiency has increased by an order and the numerical convergence depending on the initial iteration has greatly improved.The subject supported by the Chinese Foundation of Aeronautical Science     

线性加指数效用函数的组合投资研究

应用无差异曲线法较好地解决了在投资决策分析中具有特殊重要地位的线性加指数效用函数的最优组合投资比例的求解问题,并给出实例予以说明.     

On the field of definition for modularity of CM elliptic curves

The purpose of this paper is to decide the conditions under which a CM elliptic curve is modular over its field of definition.     

关于曲线积分中值定理“中间点”的一个一般性结果

讨论了第一类曲线积分中值定理“中间点”的渐近性质,得到了更具一般性的新结果.     

COSMIC(90): An improved molecular mechanics treatment of hydrocarbons and conjugated systems

Summary Four modifications to the COSMIC molecular mechanics force field are described, which greatly increase both its versatility and the accuracy of calculated conformational energies. The Hill non-bonded van der Waals potential function has been replaced by a two-parameter Morse curve and a new H-H potential, similar to that in MM3, incorporated. Hydrocarbon energies in particular are much improved.A simple iterative Hückel pi-electron molecular orbital calculation allows modelling of conjugated systems. Calculated bond lengths and rotational barriers for a series of conjugated hydrocarbons and nitrogen heterocycles are shown to be as accurate as those determined by the MM2 SCF method.Explicit hydrogen-bonding potentials for H-bond acceptor-donor atom pairs have been included to give better hydrogen bond energies and lengths. The van der Waals radii of protonic hydrogens are reduced to 0.5 Å and the energy well depth is increased to 1.0 kcal mol^-1.Two new general atom types, N⁺ _sp ² and O^- _sp ³ , have been introduced which allow a wide variety of charged conjugated systems to be studied. A minimum of parameterisation is required, as the new types are easily included in the Hückel scheme which automatically adjusts bond and torsional parameters according to the defined bond-order relationships.     

The Two-Point Codes with the Designed Distance on a Hermitian Curve in Even Characteristic

In Homma M and Kim SJ [2], the authors considered two-point codes on a Hermitian curve defined over fields of odd characteristic. In this paper, we study the geometry of a Hermitian curve over fields of even characteristic and classify the two-point codes whose minimum distances agree with the designed ones.     

生长曲线模型中回归系数阵的极大似然估计的精确分布

对于生长曲线模型,基于理论发展和应用效果的考虑,本文引入了Gauss型误差．在此误差下,本文研究了模型中回归系数阵的极大似然估计的精确分布,求出了此分布的密度和特征函数．