Monte-Carlo方法求解水动力学方程的原理及实践

介绍Monte Carlo方法的基本思想及求解Poisson方程的原理 ;利用Monte Carlo方法及图论建立河网计算的随机游动模型 ,并应用于实际的大型河网中 ,取得了较好的结果 .利用Monte Carlo方法计算河网可以避免解大型的稀疏矩阵 ,计算简单、灵活     

基于可靠度的非线性随机优化

针对约束函数被随机参数非线性表达的随机非线性优化问题,采用改进均值一次二阶矩法,研究基于元件可靠度或各失效模式可靠度的结构优化的数学模型,以及将随机优化问题化为确定性优化问题的方法.算例表明,本文方法得到的优化结果能满足所需的可靠度水平,具有实用和有效的特点.     

Effects of polarisation on mechanical properties and stress corrosion cracking susceptibility of 7050 aluminium alloy

The effects of polarisation (including cathodic and anodic polarisation) on mechanical properties and stress corrosion cracking (SCC) susceptibility (I_SCC) of 7050 aluminium alloy have been investigated by means of polarisation and slow strain rate test. The results of cathodic polarisation experiments showed that the I_SCC increased with shifting negatively the polarisation potential when the cathodic potential E_C≧?1100 mV(SCE), while it decreased with shifting negatively the polarisation potential when the cathodic potential E_CI_SCC increased severely with increasing the polarisation potential. In addition, the extents for the effect of polarisation potential on I_SCC were different among the 7050 aluminium alloy under various aging states. Polarisation was of the biggest effect on the I_SCC of under aged state, the smallest effect of over aged state and the middle effect of peak aged state. The SCC mechanism of aluminium alloy was a combination of anodic dissolution and hydrogen embrittlement, and the effects of hydrogen on SCC increased with increasing the hydrogen concentration.     

油田措施配置随机动态目标规划

针对油田开发实际中存在的随机现象和动态特征,建立了油田措施的目标随机动态规划模型,并采用具有擅长全局搜索和高度鲁棒性特点的遗传算法进行求解。应用结果表明,该模型给出了各项稳产措施的增油量 、增水量及生产费用,使措施配置更能反映实际生产动态。该方法是油田开发方案调整的极为有力的工具。     

A model of finite-step random walk with absorbent boundaries

This paper proposes a model of finite-step lattice random walk with absorbent boundaries. We address a problem of optimal stop for this model, which is defined as the absorbent boundary value with maximum profit. Compared with many existing optimal stop investigations in the random process, our study only considers the small-sample behaviour (i.e., small number of steps behaviour) and does not consider the limit behaviour of the walk. The optimal stop time is given based on classical probability computation. Since the small-sample is more practical and common than the large-sample in many real world problems, the result obtained in this paper may provide some useful guidelines for real applications associated with the finite-step random walk such as the stock market and gambling games.     

The initial stage of high‐pressure induced β‐lactoglobulin aggregation: the long‐run simulation

Research objective of this study was to clarify how the initial stage of high‐pressure induced aggregation of β‐lactoglobulin takes place. For this purpose a special simulation method was developed. Distinctive features of this approach are: the lowest resolution model of protein particles, the local interaction potential and the abandonment of the continuous simulation of particle trajectories. Relatively short molecular dynamics trajectories are used only in order to find the average time step between the collisions. Results of particle collisions that occur with some probability, are separately (‘statically’) modelled using a random variable. This allows the analysis of the process which takes place within 10²–10³ s real‐time, with an existing probability of successful collision of 10^?10–10^?11. Modelling results confirm that at the initial stage of aggregation of 0.5–2% solutions with a neutral pH‐value only dimers as well as trimers arise due to SH/S–S interaction. In addition it was shown that aggregation follows the general principles of the reaction‐limited aggregation of colloids. The proposed approach could further be used in research projects examining the aggregation of β‐lactoglobulin or similar systems.     

基于双载波随机驻留策略的HSDPA RAB分配成功率优化

通过改变双载波区域UE空闲模式下的驻留策略,提出了双载波随机驻留的方式,有效地提升了HSDPA用户的接入成功率,提升了网络质量,改善用户感知。     

DiffPRFs:random forest under differential privacy

A differential privacy algorithm DiffPRFs based on random forests was proposed.Exponential mechanism was used to select split point and split attribute in each decision tree building process,and noise was added according to Laplace mechanism.Differential privacy protection requirement was satisfied through overall process.Compared to existed algorithms,the proposed method does not require pre-discretization of continuous attributes which significantly reduces the performance cost of preprocessing in large multi-dimensional dataset.Classification is achieved conveniently and efficiently while maintains the high accuracy.Experimental results demonstrate the effectiveness and superiority of the algorithm compared to other classification algorithms.     

Combining random generators by group operation

In this article, we show that combining random generators by group operation improves the statistical properties of the composite. It gives an effective way of finding random generators more and more close to the uniform. Moreover, we obtain an effective estimation of the speed of convergence to the uniform generator.     

Evidence for compliance controlled oxygen vacancy and metal filament based resistive switching mechanisms in RRAM

We present electrical evidence on asymmetric metal-insulator-semiconductor (MIS) based test structures in support of the presence of two different independent switching mechanisms in a resistive random access memory (RRAM) device. The valid mechanism for switching depends on the compliance capping (I_gl) for forming/SET transition. Our results convincingly show that low compliance based switching only involves reversible oxygen ion drift to and from oxygen gettering gate electrodes, while high compliance switching involves formation and rupture of conductive metallic nanofilaments, as verified further by our physical analysis investigations. We have observed this unique dual mode switching mechanism only in NiSi-based gate electrodes, which have a moderate oxygen solubility as well as relatively low melting point.