Implications of genetic variation within and between human populations

Human variation is extensive both within and between populations. This variation affects all traits, including the susceptibility to disease and responses to the environment. For social and political reasons, we tend to think of our species or our own national population as being divided into meaningful groups variously called by such terms as races, or ethnic groups, and there is often an implicit assumption that these groups are relatively homogeneous within and very different between. Examples are given to show that there are major differences in the frequency of genetic susceptibility to various diseases between some such groups, but that the amount of variation among individuals is extensive even within such groups. Population subdivision can be useful for studying disease, but human variation is a general phenomenon tied to groups via their population and geographic history, not value-based categorical differences.     

Fast algorithms for computing the tripartition-based distance between phylogenetic networks

Consider two phylogenetic networks and ’ of size n. The tripartition-based distance finds the proportion of tripartitions which are not shared by and ’. This distance is proposed by Moret et al. (2004) and is a generalization of Robinson-Foulds distance, which is orginally used to compare two phylogenetic trees. This paper gives an -time algorithm to compute this distance, where h is the number of hybrid nodes in and ’ while k is the maximum number of hybrid nodes among all biconnected components in and ’. Note that $k \ll h \ll n$ in a phylogenetic network. In addition, we propose algorithms for comparing galled-trees, which are an important, biological meaningful special case of phylogenetic network. We give an $O(n)$-time algorithm for comparing two galled-trees. We also give an -time algorithm for comparing a galled-tree with another general network, where h and k are the number of hybrid nodes in the latter network and its biggest biconnected component respectively.     

一种求解TSP问题的改进遗传算法

TSP问题是一类经典的NP完全组合优化问题，传统的优化方法由于计算复杂性过大而难以求得全局最优解。遗传算法等智能优化算法在求解这类组合优化问题中表现出了强劲的潜力。作者利用遗传算法对TSP问题进行了研究分析，针对几组benchmark数据进行了仿真实验，在实验的基础上探索了遗传算子和遗传参数的优化设计，实验结果证明了遗传算法在解决TSP问题上的可行性和有效性。     

一种定向交叉的单纯形遗传算法

针对交叉算子产生个体的无方向性特点,提出了一种改进的实数交叉算子.该交叉算子吸收了单纯形法的思想,每次交叉都朝着当代种群最优解的方向进行,以提高算法的局部搜索能力.为了防止算法陷入局部极值,还提出了一种爬坡算子.实验证明,改进算法在提高收敛速度、摆脱局部最优解方面极为有效.     

Rapid evolution of larval development time during mass-rearing in the melon fly, Bactrocera cucurbitae

A quantitative genetic analysis of rapid evolution of a life history trait has been conducted on the first 24 generations of mass-rearing in the melon fly Bactrocera cucurbitae (Coquillett) (Diptera: Tephritidae). The phenotypic variance of larval development time in each generation was divided into genetic and residual components. Mean and phenotypic coefficients of variation of larval development time decreased gradually as generations proceeded as a result of artificial selection for shorter larval period in the mass-rearing procedure. There was a trend that additive genetic coefficients of variation in larval development time decreased with generations. These changes are entirely attributed to genetic responses to laboratory selection under the mass-rearing environment because the population was maintained at a very large size so as to exclude random genetic drift and inbreeding depression, which would be other factors responsible for the observed genetic changes. The residual coefficients of variation in larval development time did not change with generations. Realized heritability of larval development time was low. The heritabilities for larval development time estimated from parent–offspring regression at generations 60 and 70, when the evolutionary plateau was asymptotically reached, were not significantly larger than 0. Received: April 22, 1999 / Accepted: September 20, 1999     

A hybrid approach to constrained evolutionary computing: Case of product synthesis

Evolutionary computing (EC) is comprised of techniques involving evolutionary programming, evolution strategies, genetic algorithms (GA), and genetic programming. It has been widely used to solve optimization problems for large scale and complex systems. However, when insufficient knowledge is incorporated, EC is less efficient in terms of searching for an optimal solution. In addition, the GA employed in previous literature is modeled to solve one problem exactly. The GA needs to be redesigned, at a cost, for it to be applied to another problem. Due to these two reasons, this paper develops a generic GA incorporating knowledge extracted from the rough set theory. The advantages of the proposed solution approach include: (i) solving problems that can be decomposed into functional requirements, and (ii) improving the performance of the GA by reducing the domain range of initial population and constraining crossover using the rough set theory. The solution approach is exemplified by solving the problem of product synthesis, where there is a conflict between performance and cost. Manufacturing or assembling a product of high performance and quality at a low cost is critical for a company to maximize its advantages. Based on our experimental results, this approach has shown great promise and has reduced costs when the GA is in processing.     

孙中山生态思想的成因分析

孙中山作为20世纪中国的爱国巨匠,不仅有系统的革命思想,也有相当丰富而深刻的生态思想。他从中国国情出发,形成了自己独具个性特色的生态思想。     

不同固定建设费用轴——辐中心问题的一个有效算法

轴——辐网络中心问题是选址领域中已被广泛研究的问题之一,其中包括对中心选址和设计中心网络结构两方面。本文提出了一种基于遗传算法的混合启发式算法来解决具有不同建设费用的无约束中心选址问题。该方法依据中心点数与全部点数的比例上限确定中心点个数。本文提出了以流量、平均距离和建设费用为参数的确定各个点被选为中心点的指标因子i作为选址指标,并首次使用单亲遗传算法求解该问题。该算法的计算结果表明它可以有效的得到小规模不同建设费用轴—辐网络问题的最优解,并且表明当折扣系数a较小时,本文提出的混合启发式算法优于文献【13】提出的遗传算法。     

大数据下张量充分降维方法及其应用研究

在大数据时代,金融学、基因组学和图像处理等领域产生了大量的张量数据。Zhong等(2015)提出了张量充分降维方法,并给出了处理二阶张量的序列迭代算法。鉴于高阶张量在实际生活中的广泛应用,本文将Zhong等(2015)的算法推广到高阶,以三阶张量为例,提出了两种不同的算法:结构转换算法和结构保持算法。两种算法都能够在不同程度上保持张量原有结构信息,同时有效降低变量维度和计算复杂度,避免协方差矩阵奇异的问题。将两种算法应用于人像彩图的分类识别,以二维和三维点图等形式直观展现了算法分类结果。将本文的结构保持算法与K-means聚类方法、t-SNE非线性降维方法、多维主成分分析、多维判别分析和张量切片逆回归共五种方法进行对比,结果表明本文所提方法在分类精度方面有明显优势,因此在图像识别及相关应用领域具有广阔的发展前景。     

汇率预测及其经济基本面：基于多元自适应可变窗算法的构建

在人民币国际化不断推进、人民币汇率双向波动加强的背景下，构建具有优良预测能力的人民币汇率预测模型意义重大。参数模型对汇率预测的能力不仅取决于模型设定是否正确，同时还取决于能否迅速探测参数的结构性变化以使用最佳信息估计模型参数。本文构建了多元自适应可变窗算法以及时监测模型参数的时变特征，探测最大化参数同质区间。结果显示：①在中长期（3至24个月）的美元、欧元、英镑和日元兑人民币汇率的样本外推预测中，多元自适应可变窗算法能显著优于随机游走模型、购买力平价模型、弹性货币模型、利率平价模型、泰勒规则模型与偏移型泰勒规则模型这六种汇率预测主流模型，其预测能力也显著优于实时窗宽选择算法与自回归模型；在美元兑人民币汇率中长期（3至24个月）预测中，其预测误差MAE度量相比于次优模型能降低 25%～50%。②多元自适应可变窗算法能迅速捕捉美元、欧元、英镑和日元兑人民币汇率的拐点，预测人民币汇率走向并刻画人民币汇率的周期性变化，其长期（向前9个月）方向性趋势样本外推预测精度比次优模型提高了16%～40%。③断点前后的汇率动态结构性变化显示“811”汇改促进了经济基本面对汇率预期重要性的显著提升与市场风险偏好的转变。“811”汇改之后，人民币汇率预期更易受外部冲击影响。加速利率市场化建设、提高国内收入、稳定物价、坚持带管制的浮动汇率制度与有效的资本管制相结合等措施对促进汇率市场化、防止汇率风险具有重要意义。