选择模型的理论扩展分析

选择模型为解释指数律、幂律的成因提供了基于偏好的概率分析方法.通过对选择试验的研究,在视野规律和选项分布之间导出了一对变换,称其为偏序变换.对偏序变换连续化,导出了对应的连续变换.仿照离散分布,定义了连续分布的比较函数,并沟通了离散和连续分布的伴随关系.通过选择模型,对幂律访问现象进行了仿真.选择理论对于复杂巨系统和社会行为的研究具有较重要的借鉴作用,可扩展统计物理学和概率论的理论基础. 关键词： 选择模型 幂律现象 偏序变换 比较函数     

基于选择模式的幂律生成机制

自然界与社会生活中存在多种性质迥异的幂律分布现象,因而对它们的研究具有广泛而深远的意义．研究了一类由于人类的选择行为而产生的幂律现象,提出了人类选择行为的两个特性：理性与自利．据此,通过小球选择试验构建了模型．研究发现,这类幂律现象中随机变量服从一种近似于Zipf分布的分布．此分布为首次提出的一种新分布,将它命名为偏序分布．网络中已有的统计数据表明,偏序分布比Zipf分布更真实的反应了网络中统计变量服从的分布．文中对偏序分布进行了理论分析,并通过约束条件的改变,生成了Yule分布;表述了Yule分布的物理意义,并构造了比Yule分布更具现实意义的分布．说明了选择理论在解释幂律现象方面的普遍适用性及模型的可扩展性．文中在幂律现象的成因上支持还原论,即简单的人类选择行为导致了普遍的幂律现象． 关键词： 选择行为 幂律分布 偏序分布 Yule分布     

关于排队模型中任务完成所需时间的讨论

为了更好的模拟和解释人类的活动,针对优先权决定机制模型中任务完成所需时间展开讨论,分别分析了任务完成所需时间满足幂律分布和均匀分布两种情况下任务等待时间的分布规律。数值模拟显示任务等待时间的幂律分布不是来源于任务完成所需时间,而是来源于优先权决定的机制,从而验证了原模型中关于任务完成所需时间假设的合理性。     

非均齐超网络中标度律的涌现—–富者愈富导致幂律分布吗?

建立非线性择优连接非均齐超网络演化模型,研究非均齐超网络演化机制和拓扑性质.使用Poisson过程理论和连续化方法对模型进行分析,给出超网络超度的特征方程.利用超度特征方程不仅证明网络稳态平均超度分布存在,而且获得超度分布的解析表达式.分析表明这个网络具有"富者愈富"现象.仿真实验和理论分析相符合.随着网络规模的增大,这个动态演化的非均齐超网络的超度分布表现出拉直指数分布的特征,而不一定是幂律分布.结果表明"富者愈富"不一定导致幂律分布.     

幂律分布研究简史

自然界与社会生活中存在各种各样性质迥异的幂律分布现象,因而对它们的研究具有广泛而深远的意义.近年来,借助于有效的物理和数学工具以及强大的计算机运算能力,科学家们对幂律分布的本质有了进一步深层次的理解.文章从统计物理学的角度,简要介绍了幂律分布的研究史以及最新的进展,并对它的形成机制及动力学影响作了一些简要的阐述.     

基于平均场理论的微博传播网络模型

微博是在通过用户关注机制建立的用户网络上分享实时信息的社交平台,而微博消息主要通过用户的转发行为使消息在用户网络上传播.掌握微博消息的传播机制,对研究微博上舆论谣言的传播、产品推广等具有指导作用.本文通过对微博传播网络的结构分析来探索微博传播过程,利用新浪微博数据,建立微博传播网络,分析该网络的生成机制,使用平均场论的方法,推导微博传播网络的度分布模型.实验结果表明:微博传播网络的度分布是时间相依的,在特定时间下网络的度分布服从幂律分布.     

新节点的边对网络无标度性影响

分析新节点边对网络无标度性的影响.虽然亚线性增长网络瞬态平均度分布尾部表现出了幂律分布性质，但是，这个网络的稳态度分布并不是幂律分布，由此可见，计算机模拟预测不出网络稳态度分布，它只能预测网络的瞬态度分布.进而建立随机增长网络模型，利用随机过程理论得到了这个模型的度分布的解析表达式，结果表明这个网络是无标度网络.     

扩展幂律型脱钉和高温超导体及MgB2的电阻转变

近来,关于无序涡旋物质凝聚体的著名的涡旋玻璃理论受到了来自某些新实验的严重质疑.我们重新审查了高温超导体的临界标度,发现超导混合态的直流伏安特性具有扩展幂律的一般形式.由该扩展幂律型非线性方程模拟的等温线和Strachan等人得到的实验数据[Phys. Rev. Lett. 87(2001), 067007]符合得很好.文中还讨论了长期以来围绕纵向电阻率和霍尔电阻率的临界指数的某些争议.     

增长速度对合作网络参与者节点度分布的影响

采用率方程对合作网络自组织演化模型进行解析,得到参与者节点度分布服从Yule-Simon分布,这种分布可以用漂移幂律分布近似.分析参与者节点增长速度与参与者节点度分布之间的关系,发现随节点增长速度加快,参与者节点度分布远离幂律,表明节点增长速度是除优先连接之外的另一个影响合作网络拓扑性质的重要因素.通过对城市公交网络和科研合作网络的实证研究,验证了度分布解析结果的正确性.结合实证研究,探讨了合作网络中参与者节点增长速度的形成机制及其实际意义.     

一种新型电力网络局域世界演化模型

现实世界中的许多系统都可以用复杂网络来描述,电力系统是人类创造的最为复杂的网络系统之一.当前经典的网络模型与实际电力网络存在较大差异.从电力网络本身的演化机理入手,提出并研究了一种可以模拟电力网络演化规律的新型局域世界网络演化模型.理论分析表明该模型的度分布具有幂尾特性,且幂律指数在3—∞之间可调.最后通过对中国北方电网和美国西部电网的仿真以及和无标度网络、随机网络的对比,验证了该模型可以很好地反映电力网络的演化规律,并且进一步证实了电力网络既不是无标度网络,也不是完全的随机网络. 关键词： 电力网络 演化模型 局域世界 幂律分布     

Universality in solar flare and earthquake occurrence

Earthquakes and solar flares are phenomena involving huge and rapid releases of energy characterized by complex temporal occurrence. By analyzing available experimental catalogs, we show that the stochastic processes underlying these apparently different phenomena have universal properties. Namely, both problems exhibit the same distributions of sizes, interoccurrence times, and the same temporal clustering: We find after flare sequences with power law temporal correlations as the Omori law for seismic sequences. The observed universality suggests a common approach to the interpretation of both phenomena in terms of the same driving physical mechanism.     

Antibunching and blinking from a single colloidal CdSe quantum dot

We have investigated the optical properties of single CdSe/ZnS nanocrystals by conducting combinations of experiments on antibunching and photoluminescence intermittence under different experimental conditions.Based on photoluminescence in an antibunching experiment,we analyzed the emission lifetime of QDs by using stretched exponentials.The difference between the parameters obtained from average lifetimes and stretched exponents were analyzed by considering the effect of nonradiative emission.An Auger-assisted tunneling model was used to explain the power law exponents of off time distribution.The power law exponent under high excitation power was correlated with a higher Auger ionization rate.Using the parameters obtained from stretched exponential function and power law,the antibunching phenomena at different time and under different excitation intensity were analyzed.     

Are dragon-king neuronal avalanches dungeons for self-organized brain activity?

Recent experiments have detected a novel form of spontaneous neuronal activity both in vitro and in vivo: neuronal avalanches. The statistical properties of this activity are typical of critical phenomena, with power laws characterizing the distributions of avalanche size and duration. A critical behaviour for the spontaneous brain activity has important consequences on stimulated activity and learning. Very interestingly, these statistical properties can be altered in significant ways in epilepsy and by pharmacological manipulations. In particular, there can be an increase in the number of large events anticipated by the power law, referred to herein as dragon-king avalanches. This behaviour, as verified by numerical models, can originate from a number of different mechanisms. For instance, it is observed experimentally that the emergence of a critical behaviour depends on the subtle balance between excitatory and inhibitory mechanisms acting in the system. Perturbing this balance, by increasing either synaptic excitation or the incidence of depolarized neuronal up-states causes frequent dragon-king avalanches. Conversely, an unbalanced GABAergic inhibition or long periods of low activity in the network give rise to sub-critical behaviour. Moreover, the existence of power laws, common to other stochastic processes, like earthquakes or solar flares, suggests that correlations are relevant in these phenomena. The dragon-king avalanches may then also be the expression of pathological correlations leading to frequent avalanches encompassing all neurons. We will review the statistics of neuronal avalanches in experimental systems. We then present numerical simulations of a neuronal network model introducing within the self-organized criticality framework ingredients from the physiology of real neurons, as the refractory period, synaptic plasticity and inhibitory synapses. The avalanche critical behaviour and the role of dragon-king avalanches will be discussed in relation to different drives, neuronal states and microscopic mechanisms of charge storage and release in neuronal networks.     

Robust feed-back control of travelling waves in a class of reaction-diffusion distributed biological systems

Reaction-Diffusion (RD) mechanisms can describe many biological phenomena such as neuron firing in the brain, the heartbeat, cellular organization activities or even biological disorders such as fibrillation. The FitzHugh-Nagumo (FHN) model is a particular case of RD systems. It is able to capture the key features of many biological processes and since it is relatively simple it has been widely employed during recent years. Some examples of its predictive capabilities include the representation of the normal behavior of some physiological phenomena, related to a travelling plane wave, as well as biological disorders associated with spiral or irregular fronts. The objective of this work is to design a control law that is able to stabilize complex behaviors (travelling plane wave) in biological systems using the FHN model as a case study. Since, in biological systems there usually exists a lack of detailed information on the system structure, our control law will be designed to be robust, i.e., it must be able to reach the predefined reference regardless the presence of structural uncertainties. To this purpose, we will extend some classical results on the finite-dimensional robust control theory to RD systems by means of order reduction techniques, in particular the Proper Orthogonal Decomposition method.     

夫琅禾费衍射的计算机模拟

根据夫琅禾费衍射理论,利用数学软件及计算机蒙特卡罗随机抽样办法,准确地再现了现象及规律,可应用于辅助教学。     

Power law behavior related to mutual synchronization of chemically coupled map neurons

The widely represented network motif, constituting an inhibitory pair of bursting neurons, is modeled by chaotic Rulkov maps, coupled chemically via symmetrical synapses. By means of phase plane analysis, that involves analytically obtaining the curves guiding the motion of the phase point, we show how the neuron dynamics can be explained in terms of switches between the noninteracting and interacting map. The developed approach provides an insight into the observed time series, highlighting the mechanisms behind the regimes of collective dynamics, including those concerning the emergent phenomena of partial and common oscillation death, hyperpolarization of membrane potential and the prolonged quiescence. The interdependence between the chaotic neuron series takes the form of intermittent synchronization, where the entrainment of membrane potential variables occurs within the sequences of finite duration. The contribution from the overlap of certain block sequences embedding emergent phenomena gives rise to the sudden increase of the parameter characterizing synchronization. We find its onset to follow a power law, that holds with respect to the coupling strength and the stimulation current. It is established how different types of synaptic threshold behavior, controlled by the gain parameter, influence the values of the scaling exponents.     

Fragmentation of shells

We present a theoretical and experimental study of the fragmentation of closed thin shells made of a disordered brittle material. Experiments were performed on brown and white hen egg shells under two different loading conditions: impact with a hard wall and explosion by a combustible mixture. Both give rise to power law fragment size distributions. A three-dimensional discrete element model of shells is worked out. Based on simulations of the model, we give evidence that power law fragment mass distributions arise due to an underlying phase transition which proved to be abrupt for explosion and continuous for impact. We demonstrate that the fragmentation of closed shells defines a new universality class of fragmentation phenomena.