用矩匹配方法计算非线性风险度

矩匹配方法是用来求解非线性风险度(Value at Risk，简称：VaR)的一种普遍性方法，它是先假定样本经验分布服从已知分布族，然后运用矩匹配估计方法估计相应的参数，得到资产回报样本的密度函数，再计算风险度VaR；本采用的Johnson分布族是矩匹配方法的直接应用，并且计算出来的结果与局部Monte．Carlo结果进行了比较。并通过实证分析认为这种方法是一种良好的计算非线性VaR方法。     

我国外汇风险管理中的内部模型选择

在VaR计算中，一般都有市场变量的日回报服从正态分布的假定，但汇率的变化呈非正态分布。随着我国加入了WTO后金融市场的进一步开放，选择适合我国外汇风险管理需要的统计模型已极为紧迫。本提出了一种用有偏度的混合正态分布计算VaR的方法，使市场风险管理可不受市场变量正态性假定的限制，其参数通过分位点到分位点的对应，参照GARCH模型进行更新，从而可以对加报分布的左右胖尾进行解释。本中选取了八种与我国经济密切相关的外汇，分别计算其参数ξ、q、α和β的最优值。运算结果显示，有偏度的混合正态分布对货币组合的模拟大大优于对个体货币的拟合。     

VaR测度下的风险对冲策略研究

本文分别在正态分布和任意分布设定下讨论最小在险价值(VaR)的风险对冲问题。在正态分布设定下,本文深入讨论最小方差对冲比率和最小VaR对冲比率的性质,并得出最小VaR对冲策略下组合收益率的均值和方差大于最小方差策略下组合收益率的均值和方差。在任意分布设定下,本文构建一种新的VaR对冲模型,该模型引入非参数核估计方法对VaR进行估计,然后基于VaR核估计量建立风险对冲问题,实现风险估计与风险对冲同步进行。实证结果非常稳健地表明,不做任何分布假设下的核估计法得到的风险对冲效果优于最小方差对冲策略和正态分布设定下的最小VaR对冲策略。     

极值理论在风险度量中的应用--基于上证180指数

精确度量风险是金融风险管理的关键问题。本引入广义帕雷托分布代替传统的正态分布等，精确描述金融收益的厚尾特征。并将基于广义帕雷托分布的VaR模型和其它模型方法，如GARCH(1，1)、GARCH(1，1)-t、历史模拟法、方差-协方差方法，进行比较分析。实证研究表明，基于广义帕雷托分布的VaR模型比传统的模型方法更适合厚尾分布高分位点的预测，并且其预测结果比较稳定。这使得基于广义帕雷托分布的VaR模型成为VaR度量方法中最稳健的方法之一。     

基于Copula的VsR度量与事后检验

估计VaR的传统方法有三种:协方差矩阵法、历史模拟法和蒙特仁洛模拟法。通常,文献中认为刚蒙特卡洛模拟法度量VaR有很多方面的优点。但是,本文通过实证检验发现,使用传统蒙特卡洛模拟法估计的VaR偏小,事后检验效果很不理想。本文引入Copula函数来改进传统的蒙特卡洛模拟法。Copula函数能将单个边际分布和多元联合分布联系起来,能处理非正态的边际分布,并且它度量的相关性不再局限于线性相关性。实证检验表明,基于Copula的蒙特卡罗模拟法可以更加准确地度量资产组合的VaR。     

对数正态分布的VAR数学模型及其计算

VaR(Value at Risk)是一种以规范的统计技术来度量市场风险的新标准,目前在金融数学领域被广泛使用,它是指在正常的市场条件和给定的置信度下,在给定的持有期间内,测度某一投资组合所面临的最大的潜在损失的数学方法.传统的VaR计算方法在计算开放式基金时,可能存在着低估风险的情况.着重论述了VaR模型的数学原理以及该模型的计算方法,运用对数正态分布假设来评估开放式基金的风险,以验证其结果是否更加接近实际风险值.     

基于SVt-EVT-Vine Copula的投资组合VaR研究

针对现有风险度量模型不能准确的模拟高维金融资产收益率风险,以上证指数、沪深300指数和股指期货指数为例,首先利用SVt和EVT对各序列的边缘分布进行建模,然后采用Vine Copula方法分析多序列之间的秩相关关系和极大似然值估计法估计参数,得到RVine,CVine和DVine三种不同树结构的分解模型,通过Monte Carlo模拟法计算出在同一边缘分布不同Vine Copula方法下和在不同边缘分布同一Vine Copula方法下单资产和投资组合的金融风险VaR.经实证检验并分析对比,VaR和返回式检验均表明SVt和EVT相结合对边缘分布有较好的拟合效果,再运用RVine描述资产间的相依结构在度量投资组合金融风险方面更准确合理.     

基于Copula-GJR-Skewt模型的投资组合风险预测研究

针对多元投资组合的风险预测,采用GJR-Skewt模型刻画单资产的厚尾、有偏特征,以及Copula模型刻画多元投资组合的非线性相关结构,用Monte Carlo方法模拟金融资产的随机分布,并结合滚动时间窗法,对投资组合的未来风险进行样本外动态预测.实证结果表明,Copula-GJR-Skewt模型对资产收益的风险预测能取得满意的效果;在VaR预测性能上,以GJR-Skewt模型作为边缘分布函数时,即使存在系统偏差,也能取得最优预测结果;预设残差服从有偏学生分布时,VaR的预测结果优于正态分布;传统的Garch-Guassian模型预测能力最差.     

资产组合非等间隔日内在险价值研究

当前对资产组合在险价值(VaR)的研究仅限于等间隔抽样数据的建模。本文提出资产组合的非等间隔日内在险价值(Irregularly Spaced Intraday Value at Risk,ISIVaR)研究方法,克服资产组合逐笔交易数据非等间隔且不同步问题,利用逐笔交易数据所包含的丰富市场微观结构信息对VaR进行估计。该方法基于更新时间方法将非同步的资产组合标值序列同步化;运用Copula理论建立资产组合的非等间隔日内波动率模型,并捕捉资产组合中各资产在截面上的相关关系;最后利用这种截面相关关系,使用蒙特卡洛模拟技术估计出资产组合的ISIVaR。实证部分利用真实的逐笔交易数据验证了上述方法的有效性。     

基于有效Monte Carlo模拟的外汇期权组合非线性VaR模型

为了克服极小概率事件发生概率估计的困难,提出了把重要抽样技术发展到外汇期权组合非线性VaR模型中,估计出组合损失概率。为了进一步达到减少模拟估计误差目的,在重要抽样技术基础上使用分层抽样技术,进行更有效的Monte Carlo模拟。数值结果表明,重要抽样技术算法比常用Monte Carlo模拟法的计算效率更有效;而重要抽样技术和分层抽样技术相结合算法比重要抽样技术算法更有效地减少模拟所要估计的组合损失概率的方差,有着更高的计算效率。     

A successful algorithm for solving directed Hamiltonian path problems

In this paper we present a graph-theoretic polynomial algorithm which has positive probability of finding a Hamiltonian path in a given graph, if there is one; if the algorithm fails, it can be rerun with a randomly chosen starting solution, and there is again a positive probability it will find an answer. If there is no Hamiltonian path, the algorithm will always terminate with failure. We call this a Successful Algorithm because it has high (close to 1) empirical probability of success and it works in polynomial time. Some basic theoretical results concerning spanning arborescences of a graph are given. The concept of a ramification index is defined and it is shown that ramification index of a Hamiltonian path is zero. The algorithm starts with finding any spanning arborescence and by suitable pivots it endeavors to reduce the ramification index to zero. Probabilistic properties of the algorithm are discussed. Computational experience with graphs up to 30 000 nodes is included.     

Combining the Scalability of Local Search with the Pruning Techniques of Systematic Search

Systematic backtracking is used in many constraint solvers and combinatorial optimisation algorithms. It is complete and can be combined with powerful search pruning techniques such as branch-and-bound, constraint propagation and dynamic variable ordering. However, it often scales poorly to large problems. Local search is incomplete, and has the additional drawback that it cannot exploit pruning techniques, making it uncompetitive on some problems. Nevertheless its scalability makes it superior for many large applications. This paper describes a hybrid approach called Incomplete Dynamic Backtracking, a very flexible form of backtracking that sacrifices completeness to achieve the scalability of local search. It is combined with forward checking and dynamic variable ordering, and evaluated on three combinatorial problems: on the n-queens problem it out-performs the best local search algorithms; it finds large optimal Golomb rulers much more quickly than a constraint-based backtracker, and better rulers than a genetic algorithm; and on benchmark graphs it finds larger cliques than almost all other tested algorithms. We argue that this form of backtracking is actually local search in a space of consistent partial assignments, offering a generic way of combining standard pruning techniques with local search.     

Classification of 8-Dimensional Trilinear Alternating Forms over GF(2)

Let V be an n-dimensional vector space over a finite field and let f be a trilinear alternating form over V. For such forms we introduce a new invariant called radical polynomial and investigate its behaviour, in particular in the case of the 2-element field. We show that it is compatible with direct products of forms and how it is related to its values on dimension n ? 1. Moreover, it turns out that it is full up to dimension 7. On the other hand, on higher dimensions it is no more full and it is necessary to generalize it to obtain (using computer) a classification of forms on dimension 8 over the 2-element field. This classification is provided, together with the sizes of stabilizers of the corresponding forms.     

AHP中判断矩阵的满意的一致性的研究

对层次分析法(AHP)中的判断矩阵,给出了它具有满意的一致性的充分必要条件,也就是,给出了检验它是否有满意的一致性的一种方法,在否定的情况下,还给出了修正它的一种方法使得它的每个主子矩阵都有满意的一致性.     

Fuzzy集的基数

关于Fuzzy集的基数，[1]中曾对有限支集的Fuzzy集以及极苛刻条件下的无限Fuzzy集给出过一种定义。但是，正如本文将要指出的那样，该定义是不合理的。本文将从研究Fuzzy映射入手，给出Fuzzy集基数的一般性定义。基于这一定义，不但得到有关基数的大部分结论，而且有其自身的特殊性。     

Neo-Mooreanism,contextualism, and the evidential basis of scepticism

Two of the main forms of anti-scepticism in the contemporary literature—namely, neo-Mooreanism and attributer contextualism—share a common claim, which is that we are, contra the sceptic, able to know the denials of sceptical hypotheses. This paper begins by surveying the relative merits of these views when it comes to dealing with the standard closure-based formulation of the sceptical problem that is focussed on the possession of knowledge. It is argued, however, that it is not enough to simply deal with this version of the sceptical challenge, since there is a more fundamental sceptical problem underlying the standard closure-based sceptical argument that can be expressed in terms of the evidential basis of our beliefs. Whilst it is argued that neo-Mooreanism has a slight edge over attributer contextualism when it comes to dealing with the closure-based formulation of the sceptical problem, it is claimed that this view is in an ever stronger dialectical position when it comes to the more pressing evidential formulation of the sceptical problem. It is shown that this is so even if one adapts the attributer contextualist thesis along the lines suggested by Michael Williams and Ram Neta so that it is explicitly designed to deal with the evidential variant of the sceptical problem.     

回路1、2-弦图的逆M矩阵完备及其算法设计

对已定元均不为零的部分逆M矩阵,通过变换使其对角线上元素均为1后,根据其所对应图形的特点,得到结果如下:(a)若其所对应图形为简单有向回路或回路1-弦图,具有逆M矩阵完备式当且仅当所有简单有向回路的回路积均小于1.(b)若其所对应图形为回路2-弦图,具有逆M矩阵完备式当所有简单有向回路满足回路积小于1,且对其中依次在两个顶点处相交的有向回路标明层次后,任一有向回路的回路积均小于与其相连接的上一层的有向回路的回路积.     

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。     

Starlike approximations of analytic functions

When an analytic function is not univalent, it is often of interest to approximate it by univalent functions. In this paper we introduce a measure of the non-univalency of a function and we derive a method for constructing the best starlike univalent approximations of analytic functions with respect to it, suitable for both practical problems and numerical implementation.     

The Existence of Almost Periodic Solution and Periodic Solutions

In this paper, it is obtained that a periodic system has an almost periodic solution if it has a solution x=\phi(t) uniformly stable with respect to \Omega_\phi ,and has a periodic solution if x=\phi(t) is weakly uniformly asymptotically stable with respect to \Omega_\phi. Meanwhile, it is also obtained that a uniformly almost periodic system has an almost periodic solution if it has a solution x=\phi(t) uniformly asymptotically stable with respect to A_\phi^j.