应用结构异质性阈值模型定位阈值性状基因位点 : 仿真研究(英文)

蒙特卡罗仿真研究表明 :对处理在基因定位阈值性状中出现的超离中趋势现象 (结构异质性 ) ,我们提出的结构异质性模型是一个高效的统计方法 .它表现在高效率的统计检验、准确的参数估计和基因定位等方面 .病态或奇异费歇信息矩阵是在基因连锁定位分析中的一个突出的算法问题 ,仿真数据显示它们的发生率可以达到 2 8% .我们提出的应用奇异根分解方法可以有效地解决这一算法问题 .对比常规阈值模型 ,结构异质性阈值模型有较高的算法稳定性 .     

应用结构异质性阈模型定位阈性状基因位点:I理论研究(英文)

结构异质性是在定位阈性状基因位点过程中普遍存在的一个问题 .例如 ,数量遗传学家通常根据连锁分子标记推断性状位点基因型 ,由于性状位点基因型的不确定性 ,一个亲折方差组分被引入阈模型 .针对这个问题 ,我们提出了一个新颖的统计模型并推广到多阈值性状 .在这篇理论文章中 ,我们给出了详细的理论推导和计算策略     

应用结构异质性阈模型定位性状基因位点：I理论研究

结构异质性是在定位阈性状基因位点过程中普遍存在的一个问题。例如,数量遗传学家通常根据锁分子标记推断性状位点的基因型,由于性状位点基因型的不确定性,一个亲折方差组分被引入阈模型,针对这个问题,我们提出了一个新颖的统计模型并推广到多阈值性状。在这篇理论文章中,我们给出了详细的理论推导和计算策略。     

SNR快速算法及基因阈值确定

在很长的DNA序列中,DFT的大量运算会影响基因识别的效率,对此提出了在Voss映射下仅依赖于碱基在三种不同位置上出现频率的信噪比计算公式,推导发现在Voss与Z-curve映射下其频谱之间存在4倍缩放关系,仿真实验验证了此结果.针对阈值判别的主观性和经验性等缺陷,运用仿真实验方法确定不同类基因的阈值.从敏感性、特异性和精确度三方面对不同阈值下的测试结果进行了评估.同时提出了基于Bootstrap重复抽样的基因最优阈值算法,对不同类基因的最优阈值进行了预测,其中人和鼠类的最优阈值为1.930,并分析了算法的有效性和可行性,其精确度达到了92.8%.     

基于l_∞-模的Hankel矩阵填充的保结构阈值算法

文章基于l_∞-范数的性质及奇异值阈值方法,提出Hankel矩阵填充的一种算法.该算法保证每次迭代产生的填充矩阵是可行的Hankel矩阵,不仅减少了奇异值分解所用的时间,而且获得更精确的填充矩阵.同时,讨论了新算法的收敛性.最后通过数值实验以及简单的图像修复证明新算法比加速邻近梯度算法、阈值的增广Lagrange乘子算法以及基于F-模的Hankel矩阵填充的保结构阈值算法更有效.     

基于迭代投影的梯度硬阈值追踪算法

梯度硬阈值追踪算法是求解稀疏优化问题的有效算法之一.考虑到算法中投影对最优解的影响,提出一种比贪婪策略更好的投影算法是很有必要的.针对一般的稀疏约束优化问题,利用整数规划提出一种迭代投影策略,将梯度投影算法中的投影作为一个子问题求解.通过迭代求解该子问题得到投影的指标集,并以此继续求解原问题,以提高梯度硬阈值追踪算法的计算效果.证明了算法的收敛性,并通过数值实例验证了算法的有效性.     

集体行动的阈值模型

主要讨论集体行动的特征,社会机理,集体行动研究的主要两类模型阈值模型和效用模型.然后,综合考虑两类模型的优缺点并引入心理阈值理论,在经典Granovetter阈值模型基础上给出了一种融合心理学阈值理论和经济学效用理论的集体行为阈值模型.并在考虑群体空间因素和朋友社会影响强度条件下对所提出的模型进行了仿真研究,得到了集体行动的相变临界值,均衡点(达到平衡状态后参与数量).文章尝试运用阈值理论解释骚乱暴动等激进群体行为,以及Web上群体观点极化与社会认同理论之间的联系.最后,讨论了提出的模型与Granovetter模型之间的不同及应用范围.     

基于改进阈值的马田系统在医疗智能诊断中的应用

人工智能与医疗数据的融合可以加速疾病诊断过程,提高诊断精度,挖掘诊断过程关键指标,改善医疗工作流程.以医疗决策者偏好为前提,提出基于二次损失函数改进阈值的马田系统算法.通过正交表与信噪比对指标进行优化,降低模型复杂度;通过对阈值的改进,提高模型的灵敏度,满足医疗工作者的决策偏好.将此方法应用于UCI乳腺癌医学数据及三甲医院哮喘临床数据,并与其他改进阈值的马田系统算法及智能算法对比分析.结果表明,改进阈值的马田系统算法识别灵敏度高,简化诊断指标,训练耗时少,是一种更为有效的医疗智能诊断方法.     

MEA-BP模型在遥感影像分类中的应用研究

遥感影像分类作为遥感技术的一个重要应用,对遥感技术的发展具有重要作用.针对遥感影像数据特点,在目前的非线性研究方法中主要用到的是BP神经网络模型.但是BP神经网络模型存在对初始权阈值敏感、易陷入局部极小值和收敛速度慢的问题.因此,为了提高模型遥感影像分类精度,提出采用MEA-BP模型进行遥感影像数据分类.首先采用思维进化算法代替BP神经网络算法进行初始寻优,再用改进BP算法对优化的网络模型权阈值进一步精确优化,随后建立基于思维进化算法的BP神经网络分类模型,并将其应用到遥感影像数据分类研究中.仿真结果表明,新模型有效提高了遥感影像分类准确性,为遥感影像分类提出了一种新的方法,具有广泛研究价值.     

稀疏正则化:Gauss-Seidel阈值迭代算法

本文考虑一类稀疏正则化问题,该类问题在机器学习、信号处理和图像处理等众多领域中被广泛研究.此类问题的一个典型特征是其诱导的阈值函数具有跳跃的不连续性.本文提出一种基于GaussSeidel的迭代算法,称作Gauss-Seidel跳跃阈值迭代算法(Gauss-Seidel iterative jumping thresholding algorithm,GSIJT),用以快速解决以上问题.本文首先证明了由GSIJT所产生序列的支撑与符号的有限收敛性.基于此收敛性质,同时利用restricted Kurdyka-Lojasiewicz(rKL)性质给出GSIJT算法的全局收敛性.此外给出了GSIJT的收敛率,并且证明了任意的极限点都是驻点.本文实施了一系列的数值实验来验证所提算法的有效性.特别地,通过与相关的阈值迭代算法进行比较,表明所提算法不仅收敛更快,同时可选择的步长范围更宽.     

模糊自组织特征映射模型

本文提出了一种模糊自组织特征映射模型算法，它将模糊ｃ－均值模型结合到Ｋｏｈｏｎｅｎ自组织算法的学习和更新策略中，从而将神经元的侧抑制作用与模糊控制策略有机地结合起来；不仅实现了关于模糊ｃ－均值的优化问题，而且还通过隶属函数的重新构造最终构成了自组织有序拓扑图。仿真结果表明收敛性得到了很大提高。     

A New Approach to Cycling in a 2-locus 2-allele Genetic Model

In the reproductive process new genetic types arise due to crossing over and recombination at the meiotic stage. A simplified biological model will be developed which incorporates this effect and the effect of selection. Although a chromosome may contain thousands of genes we will consider a simplified model consisting of two genetic loci, each containing two alleles of some gene.

The model will be then turned into a difference equation or mapping model x* = G(x,r) where x represents the frequency distribution of genotypes in a certain infinite population, x* is this distribution one generation later and r is the recombination parameter. For a certain choice of fitness and recombination parameters the mapping exhibits several fixed points. As r is varied one of the fixed points of the mapping loses its stability due to a conjugate pair of eigenvalues of the linearized mapping leaving the unit disk. It is shown that the required non-resonance conditions and “nonlinear damping” condition are satisfied and thus the fixed point undergoes a Neimark–Sacker bifurcation to a cycling or oscillatory state.

Once a cycling orbit is established one can conclude that genetic variation (over time) of the population can be maintained. This work reformulates and proves earlier observations of Alan Hastings in a way that makes the treatment of chromosomes with more genetic loci more straightforward.     

Space Mapping Approach for Particle Control in Turbulent Flows

Controlling the motion of particles in turbulent flows, the paper at hand presents an efficient space–mapping approach that is based on a hierarchy of models. The approach reduces the highly complex optimization of the k-ε turbulence model for high Reynolds–number flows (fine model) to the cheaper one of the Navier–Stokes equations for smaller Reynolds–number (laminar) flows in direct numerical simulations on coarser grids (coarse model) by help of a space–map function that maps the respective coarse model control onto the desired fine model control. The numerical results are very convincing in terms of accuracy and computational effort. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

遗传算法、差分演化和一个处理约束的映射

S.Koziel和Z.Michalewicz(1999年)提出了一个处理约束的映射,研究该映射与不同算法相结合后的不同的代数结构.从理论上证明了当其与遗传算法相结合时,该映射是同构映射,而在差分演化算法的变异操作下,该映射不是同态映射,更不是同构映射.进而表明,该映射更适宜于与遗传算法相结合,而并不太适宜于与差分演化算法(及其类似的算法)相结合。     

Application of seasonal SVR with chaotic gravitational search algorithm in electricity forecasting

Hybridization chaotic mapping functions with optimization algorithms into a support vector regression model has been shown its efficient potential to avoid converging prematurely. It is deserved to explore more possibility by hybridizing with other optimization algorithms. Electricity demand sometimes demonstrates a seasonal tendency due to complicate economic activities or climate cyclic nature. This investigation presents a SVR-based electricity forecasting model which applied a novel hybrid algorithm, namely chaotic gravitational search algorithm (CGSA), to improve the forecasting performance. The proposed CGSA employs the chaotic local search by logistic chaotic mapping function in the iteration of the original GSA to search and refine the current best solution. In addition, seasonal mechanism is also applied to deal with seasonal electricity tendency. A numerical example from an existed reference is used to illustrate the forecasting performance of the proposed SSVRCGSA model. The forecasting results indicate that the proposed model yields more accurate forecasting results than ARIMA and TF-ε-SVR-SA models.     

流通量间断守恒高阶交通流模型及其数值模拟

在非均匀道路条件下，推广了各向异性守恒高阶交通流模型(CHO模型)，获得流通量间断CHO模型，并基于其Riemann不变量性质，运用局部简化方法及δ映射算法，设计了求解流通量间断CHO模型的一阶Godunov、EO(Engquist-Osher)和LF(Lax-Friedrichs)等数值格式.通过数值模拟表明流通量间断CHO模型是合理有效的，它可以描述平衡态和非平衡态交通流，相对于流通量间断LWR(Lighthill-Whitham-Richards)模型，其能更好地刻画实际交通现象.     

A boundary preserving numerical algorithm for the Wright-Fisher model with mutation

The Wright-Fisher model is an It? stochastic differential equation that was originally introduced to model genetic drift within finite populations and has recently been used as an approximation to ion channel dynamics within cardiac and neuronal cells. While analytic solutions to this equation remain within the interval [0,1], current numerical methods are unable to preserve such boundaries in the approximation. We present a new numerical method that guarantees approximations to a form of Wright-Fisher model, which includes mutation, remain within [0,1] for all time with probability one. Strong convergence of the method is proved and numerical experiments suggest that this new scheme converges with strong order 1/2. Extending this method to a multidimensional case, numerical tests suggest that the algorithm still converges strongly with order 1/2. Finally, numerical solutions obtained using this new method are compared to those obtained using the Euler-Maruyama method where the Wiener increment is resampled to ensure solutions remain within [0,1].     

DSLC-FOA : Improved fruit fly optimization algorithm for application to structural engineering design optimization problems

In this study, we propose an improved fruit fly optimization algorithm (FOA) based on linear diminishing step and logistic chaos mapping (named DSLC-FOA) for solving benchmark function unconstrained optimization problems and constrained structural engineering design optimization problems. Based on comparisons with genetic algorithm, particle swarm optimization, FOA, LGMS -FOA, and chaotic FOA methods, we demonstrated that DSLC-FOA performed better at searching for the optimal solutions of four typical benchmark functions. The approximate optimal results were obtained using DSLC-FOA for three structural engineering design optimization problems as examples of applications. The numerical results demonstrated that the proposed DSLC-FOA algorithm is superior to the basic FOA and other metaheuristic or deterministic methods.     

A philosophy for the modelling of realistic nonlinear systems

A nonlinear dynamical system is modelled as a nonlinear mapping from a set of input signals into a corresponding set of output signals. Each signal is specified by a set of real number parameters, but such sets may be uncountably infinite. For numerical simulation of the system each signal must be represented by a finite parameter set and the mapping must be defined by a finite arithmetical process. Nevertheless the numerical simulation should be a good approximation to the mathematical model. We discuss the representation of realistic dynamical systems and establish a stable approximation theorem for numerical simulation of such systems.