RUIN PROBABILITY IN THE CONTINUOUS-TIME COMPOUND BINOMIAL MODEL WITH INVESTMENT

This article deals with the problem of minimizing ruin probability under optimal control for the continuous-time compound binomial model with investment. The jump mechanism in our article is different from that of Liu et al [4]. Comparing with [4], the introduction of the investment, and hence, the additional Brownian motion term, makes the problem technically challenging. To overcome this technical difficulty, the theory of change of measure is used and an exponential martingale is obtained by virtue of the extended generator. The ruin probability is minimized through maximizing adjustment coefficient in the sense of Lundberg bounds. At the same time, the optimal investment strategy is obtained.

A continuous time delay-difference type model(CTDDM) applied to stock assessment of the southern Atlantic albacore Thunnus alalunga

A continuous time delay-difference model(CTDDM) has been established that considers continuous time delays of biological processes.The southern Atlantic albacore(Thunnus alalunga) stock is the one of the commercially important tuna population in the marine world.The age structured production model(ASPM) and the surplus production model(SPM) have already been used to assess the albacore stock.However,the ASPM requires detailed biological information and the SPM lacks the biological realism.In this study,we focus on applying a CTDDM to the southern Atlantic albacore(T.alalunga) species,which provides an alternative method to assess this fishery.It is the first time that CTDDM has been provided for assessing the Atlantic albacore(T.alalunga) fishery.CTDDM obtained the 80%confidence interval of MSY(maximum sustainable yield) of(21 510 t,23 118 t).The catch in 2011(24 100 t) is higher than the MSY values and the relative fishing mortality ratio(F_(2011)/F_(MSY)) is higher than 1.0.The results of CTDDM were analyzed to verify the proposed methodology and provide reference information for the sustainable management of the southern Atlantic albacore stock.The CTDDM treats the recruitment,the growth,and the mortality rates as all varying continuously over time and fills gaps between ASPM and SPM in this stock assessment.

A Numerical Algorithm Based on Quadratic Finite Element for Two-Dimensional Nonlinear Time Fractional Thermal Diffusion Model

In this article,a high-order scheme,which is formulated by combining the quadratic finite element method in space with a second-order time discrete scheme,is developed for looking for the numerical solution of a two-dimensional nonlinear time fractional thermal diffusion model.The time Caputo fractional derivative is approximated by using the L2-1formula,the first-order derivative and nonlinear term are discretized by some second-order approximation formulas,and the quadratic finite element is used to approximate the spatial direction.The error accuracy O(h3+
t2)is obtained,which is verified by the numerical results.

计及调节资源AGC动态特性的离散-连续时间随机前瞻经济调度

随着具有波动性的可再生能源渗透比例不断提高,系统对调频备用的需求显著增加。为了制定更合理的调度方案和调频备用,有必要研究经济调度和自动发电控制(automatic generation control,AGC)之间的协调优化。该文提出建立考虑调节资源AGC动态特性的离散-连续时间随机前瞻经济调度模型,以制定兼顾能量平衡和调频需求的调度决策,并实现快慢资源的最优协同,增强系统的动态调节能力。为降低模型复杂度,在当前时段将调节资源的AGC动态特性建模为精细化离散模型,在前瞻时段采用简化的连续时间模型刻画调节资源的AGC动态特性,显著降低问题规模。通过在目标函数中引入合适的储能充放电成本惩罚项,将储能充放电互补约束松弛,从而消除01变量。最后,采用Benders分解算法实现并行求解,进一步降低了计算时间。该文在PJM5节点系统和IEEE118节点系统进行算例验证。结果表明,与传统模型相比,所提模型可以有效降低系统的频率偏差;与在前瞻时段考虑精细化AGC动态约束的模型相比,所提模型可显著降低计算时间。

连续时间的IS-LM模型可控制性与仿真研究

在现代经济学的标准假设下,建立连续时间的IS-LM动态模型,分析IS-LM动态模型的系统结构特征,设计连续时间的IS-LM动态模型的控制律,得到控制律的解析解并进行仿真分析。研究表明:连续时间的IS-LM动态模型具有能控性、能达性和完全能观性等系统结构特征;连续时间IS-LM动态模型系统可以任意配置其极点,使模型系统渐进稳定,达到输出调节并具有满意的动态性能,即在现代经济学的标准假设下,证实了IS-LM动态模型的可控性。根据模型假设及模型参数的经济意义设计仿真实验,仿真实验结果验证了所设计反馈控制律对连续时间IS-LM动态模型控制的有效性。因而,针对复杂多变的经济形势,科学设计经济政策,能有效实现经济调控的目标。

Demand Model for Investment-oriented Life Insurance under uncertain lifetime

In this paper, we set up continuous time model with Poisson Process to analyze demand of investment-oriented life insurance. Individual life time is assumed random, and he is received fixed income, investment-oriented life insurance is an important financial asset under this model. Dynamic programming is applied to analyze this problem. The optimal explicit solutions are obtained in the case of CRRA utilities, and draw its demand curve with numerical simulation.

Measurement,Modelling and Analysis of Residence Time Distribution Characteristics in a Continuous Hydrothermal Reactor

Understanding the residence time distribution(RTD)of a continuous hydrothermal reactor is of great significance to improve product quality and reaction efficiency.In this work,an on-line measurement system is attached to a continuous reactor to investigate the characteristics of RTD.An approach that can accurately fit and describe the experimental measured RTD curve by finding characteristic values is proposed for analysis and comparison.The RTD curves of three experiment groups are measured and the characteristic values are calculated.Results show that increasing total flow rate and extending effective reactor length have inverse effect on average residence time,but they both cause the reactor to approach a plug flow reactor and improve the materials leading.The branch flow rate fraction has no significant effect on RTD characteristics in the scope of the present work except the weak negative correlation with the average residence time.Besides,the natural convection stirring effect can also increase the average residence time,especially when the forced flow is weak.The analysis reveals that it is necessary to consider the matching of natural convection,forced flow and reactor size to control RTD when designing the hydrothermal reactor and working conditions.

Evolved differential model for sporadic graph time-series prediction

Sensing signals of many real-world network systems,such as traffic network or microgrid,could be sparse and irregular in both spatial and temporal domains due to reasons such as cost reduction,noise corruption,or device malfunction.It is a fundamental but challenging problem to model the continuous dynamics of a system from the sporadic observations on the network of nodes,which is generally represented as a graph.In this paper,we propose a deep learning model called Evolved Differential Model(EDM)to model the continuous-time stochastic process from partial observations on graph.Our model incorporates diffusion convolutional network to parameterize continuous-time system dynamics by graph Ordinary Differential Equation(ODE)and graph Stochastic Differential Equation(SDE).The graph ODE is applied to accurately capture the spatial-temporal relation and extract hidden features from the data.The graph SDE can efficiently capture the underlying uncertainty of the network systems.With the recurrent ODE-SDE scheme,EDM can serve as an accurate online predictive model that is effective for either monitoring or analyzing the real-world networked objects.Through extensive experiments on several datasets,we demonstrate that EDM outperforms existing methods in online prediction tasks.

Impact of time lags on diurnal estimates of canopy transpiration and canopy conductance from sap-flow measurements of Populus cathayana in the Qinghai–Tibetan Plateau

Recently, canopy transpiration （Ec） has been often estimated by xylem sap-flow measurements. However, there is a significant time lag between sap flow measured at the base of the stem and canopy transpiration due to the capacitive exchange between the transpiration stream and stem water storage. Significant errors will be introduced in canopy conductance （gc） and canopy transpiration estimation if the time lag is neglected. In this study, a cross-correlation analysis was used to quantify the time lag, and the sap flowbased transpiration was measured to pararneterize Jarvistype models of gc and thus to simulate Ec of Populus cathayana using the Penman-Monteith equation. The results indicate that solar radiation （Rs） and vapor pressure deficit （VPD） are not fully coincident with sap flow and have an obvious lag effect; the sap flow lags behind Rs and precedes VPD, and there is a 1-h time shift between Eo and sap flow in the 30-min interval data set. A parameterized Jarvis-type gc model is suitable to predict P. cathayana transpiration and explains more than 80% of the variation observed in go, and the relative error was less than 25%, which shows a preferable simulation effect. The root mean square error （RMSEs） between the predicted and measured Ec were 1.91×10^-3 （with the time lag） and 3.12×10^-3cm h^-1 （without the time lag）. More importantly, Ec simulation precision that incorporates time lag is improved by 6% compared to the results without the time lag, with the mean relative error （MRE） of only 8.32% and the mean absolute error （MAE） of 1.48 × 10^-3 cm h^-1.

A time fractional model to represent rainfall process

This paper deals with a stochastic representation of the rainfall process. The analysis of a rainfall time series shows that cumulative representation of a rainfall time series can be modeled as a non-Gaussian random walk with a log-normal jump distribution and a time-waiting distribution following a tempered a-stable probability law. Based on the random walk model, a fractional Fokker-Planck equation （FFPE） with tempered a-stable waiting times was obtained. Through the comparison of observed data and simulated results from the random walk model and FFPE model with tempered a-stable waiting times, it can be concluded that the behavior of the rainfall process is globally reproduced, and the FFPE model with tempered a-stable waiting times is more efficient in reproducing the observed behavior.

On iterative learning control design for tracking iteration-varying trajectories with high-order internal model

In this paper, iterative learning control （ILC） design is studied for an iteration-varying tracking problem in which reference trajectories are generated by high-order internal models （HOLM）. An HOlM formulated as a polynomial operator between consecutive iterations describes the changes of desired trajectories in the iteration domain and makes the iterative learning problem become iteration varying. The classical ILC for tracking iteration-invariant reference trajectories, on the other hand, is a special case of HOlM where the polynomial renders to a unity coefficient or a special first-order internal model. By inserting the HOlM into P-type ILC, the tracking performance along the iteration axis is investigated for a class of continuous-time nonlinear systems. Time-weighted norm method is utilized to guarantee validity of proposed algorithm in a sense of data-driven control.

Probabilistic Modelling of COVID-19 Dynamic in the Context of Madagascar

We propose a probabilistic approach to modelling the propagation of the coronavirus disease 2019 (COVID-19) in Madagascar, with all its specificities. With the strategy of the Malagasy state, which consists of isolating all suspected cases and hospitalized confirmed case, we get an epidemic model with seven compartments: susceptible (S), Exposed (E), Infected (I), Asymptomatic (A), Hospitalized (H), Cured (C) and Death (D). In addition to the classical deterministic models used in epidemiology, the stochastic model offers a natural representation of the evolution of the COVID-19 epidemic. We inferred the models with the official data provided by the COVID-19 Command Center (CCO) of Madagascar, between March and August 2020. The basic reproduction number R0 and the other parameters were estimated with a Bayesian approach. We developed an algorithm that allows having a temporal estimate of this number with confidence intervals. The estimated values are slightly lower than the international references. Generally, we were able to obtain a simple but effective model to describe the spread of the disease.

基于ResNet-ELM和迁移学习的风机齿轮箱故障诊断方法

在变工况以及混合故障情况下,采用传统的方法对风力发电机齿轮箱进行故障诊断时,存在诊断效果不佳、泛化能力差等问题,为此,提出了一种基于深度残差网络(ResNet)-极限学习机(ELM)和迁移学习(TL)的风力发电机齿轮箱故障诊断方法(TL-RN-ELM)。首先,分别介绍了连续小波变换(CWT)、卷神经网络(CNN)、深度残差网络(ResNet)、迁移学习(TL)、极限学习机(ELM)的原理;然后,基于上述原理,建立了TL-RN-ELM的风力发电机齿轮箱故障诊断模型(流程);最后,使用轴承数据集和齿轮箱数据集对上述方法进行了实验验证,即从美国凯斯西储大学(CWRU)轴承数据集和东南大学(SEU)齿轮箱数据集中进行了数据采集和处理,利用CWT将原始一维振动信号转换成了二维小波时频图像,使用CWRU轴承数据集对搭建的ResNet18模型进行了训练,生成了预训练模型;将预训练模型中的数据迁移至SEU齿轮箱数据集,微调了模块,提取了特征,并将其输入到ELM分类器,然后将分类结果与其他3类模型进行了对比分析。实验结果表明:对于从轴承到轴承、轴承到齿轮以及混合故障的小样本迁移故障诊断,TL-RN-ELM的平均准确率可达98.79%;与其他方法相比,该方法的平均准确率提升了4.73%~9.6%。研究结果表明:该方法具有良好的诊断效果和泛化能力。

基于时频域分析的车载毫米波雷达干扰抑制方法

文中针对车载调频连续波雷达之间相互干扰导致虚警和漏警的问题,提出一种在时频域基于改进经验模式分解和自回归模型相结合的干扰抑制方法。该方法首先使用经验模式分解区分出拍频信号中干扰分量主导的低阶本征模态函数,将其转换到短时傅里叶变换域后通过全局阈值方法进行干扰分量定位;其次,在时频域根据定位信息将拍频信号包含干扰的数据置零;最后,使用自回归模型对拍频信号中缺失的有用信号进行估计并插值。通过仿真和实测结果显示,该方法在精确地去除干扰分量的同时可以减少有用信号的功率损失,干扰抑制后的信号与参考信号的相关系数达到0.9697。与现有干扰抑制技术相比文中方法也体现出更优的干扰抑制性能。

中国内地GPS坐标时间序列噪声模型特征及其对站点速率影响

为了进一步评估全球定位系统(GPS)时间序列噪声模型水平和垂直速率的大小,及其对误差的影响,选取中国内地227个GPS连续基准站2010—2020年南北、东西和垂直3个方向的坐标时间序列,采用6种噪声模型或噪声组合模型对其进行噪声分析。结果表明,中国内地GPS坐标时间序列噪声模型存在多样性,且部分站点在不同方向的噪声模型也存在差异,主要以一阶高斯马尔可夫+随机漫步噪声(GGMWN)和闪烁噪声+白噪声(FNWN)为主;在100°E附近的GPS站点噪声特性差异最为显著;噪声模型与速率之间的关系分析表明噪声模型对水平向速率的大小和误差影响较小,在现实计算中可不考虑噪声对水平速度的影响,但对垂向速率的大小和误差影响显著;考虑噪声模型可有效提高垂向速率的精度,同时也可能会改变部分站点的垂向运动方向,所以在现实计算中须考虑噪声对垂向速率的影响。

Spatial-time continuous changes simulation of crop growth parameters with multi-source remote sensing data and crop growth model

本文将遥感信息与作物模型同化实现作物生长参数的时空域连续模拟,进而监测生长参数的时空域变化。首先将作物模型WOFOST(World food studies)与冠层辐射传输模型PROSAIL耦合构建WOPROSAIL模型,利用微粒群算法(PSO)通过最小化从CCD数据获取的土壤调节植被指数观测值SAVI(soil adjusted vegetation index)与耦合模型得到的模拟值SAVI’之间差值优化作物模型初始参数。通过MODIS数据反演实现参数的区域化,并将区域参数作为优化后作物模型输入参数驱动模型逐像元计算生长参数,实现生长参数的时空域连续模拟与监测,最终建立区域尺度遥感-作物模拟同化框架模型RS-WOPROSAIL。结果表明:同化模型解决了作物模型模拟空间域和遥感信息时间域的不连续问题。模型模拟的叶面积指数(LAI)、穗重(WSO)、地上总生物量(TAGP)等生长参数较好地体现了水稻生长状况时空域变化,研究区水稻模拟产量与实际产量的误差为27.4%。

H_2/Air连续旋转爆震发动机推力测试(Ⅱ)-双波模态下的推力

在环缝-喷孔对撞式喷射的H2/Air连续旋转爆震模型发动机上实现双波自持。详细分析了连续旋转爆震波以双波模态自持传播的典型波形特征和时域、频域特征。测量了模型发动机工作在双波形模态下所产生的一维推力,讨论了比冲等推力性能。时频特性和推力积分表明:出口背压为大气压时,在空气流量786.6g·s-1,氢气流量20g·s-1,当量比为0.8733的工况下,模型发动机以平均传播频率10.5809k Hz,平均传播速度1578.9m·s-1的双波模态稳定工作超过650ms。产生可靠的有效推力约808.5N。以火箭模式计算,有效排气速度为1002.3m·s-1,总比冲为102.3s;以冲压模式计算,有效排气速度(氢气消耗率)为40425m·s-1,燃料比冲为4125s,所消耗氢气的单位面积质量流率为13404g·m-2·s-1,单位推力为1027.8m·s-1。相比于单波模态,双波模态使得燃烧室内压力更为均匀,高频推力曲线振荡幅值小。爆震波头个数增多有利于推力稳定。

基于GSCPN的硬件木马攻击建模与安全性分析

针对硬件木马(hardware trojan,HT)的研究主要集中在HT设计和防御的具体策略,缺乏面向HT攻击的系统安全性量化评估模型的问题,提出了基于广义随机着色petri网(generalized stochastic coloring petri net,GSCPN)的HT攻击和安全评估模型。基于单个基本单元的HT攻击行为和状态之间的时序关系,构建了单个基本单元的HT攻击GSCPN模型,根据系统工作状态下基本单元的通信依赖关系,给出了构建系统HT攻击GSCPN模型的通用方法。求解了与GSCPN模型同构的马尔可夫链的解析解,以及采用蒙特卡罗模拟方法的仿真解,二者的误差仅为0.00183,验证了所提方法建模的准确性。通过案例分析了不同场景下系统的安全性,实验结果表明,木马攻击强度、木马恢复时间和任务执行时间均对安全性有不同程度的影响。

具有功能反应和时滞的三种群捕食-食饵扩散模型的正周期解的存在性

讨论了一类具有时滞和Michaelismenten型功能反应函数的三种群捕食-食饵扩散模型,且所有参数均依赖于时间.应用重合度连续性定理,得到了该系统正周期解存在性的充分条件.