壤中流模型研究的现状及存在问题

对国内外壤中流模型与模拟进行了较为系统的介绍,并针对这些模型提出了一种壤中流模型分类的方法,即根据模型所依据的主要原理将壤中流模型分为三大类：１）Ｒｉｃｈａｒｄｓ模型;２）动力波模型;３）贮水泄流模型．Ｒｉｃｈａｒｄｓ模型又可分为一维Ｒｉｃｈａｒｄｓ模型、二维Ｒｉｃｈａｒｄｓ模型和三维Ｒｉｃｈａｒｄｓ模型;贮水泄流模型又可分为动力贮水泄流模型和Ｂｏｕｓｉｎｅｓｑ贮水泄流模型．同时,将这３类模型进行对比,指出了它们各自优点和不足．     

墒情诊断模型的评价

对6个模型进行系统评价的目的是为模型优选提供理论、方法和参数依据。采用优选模型比例、验证方法指标、验证方式指标和异常值指标4项指标对6个独立墒情诊断模型进行了综合评价。结果是:模型优先顺序为差减统计模型、间隔天数统计模型移动统计模型比值统计模型统计模型平衡模型;逐日模型好于时段模型。合格率高的模型包括4个特征,即自变量数量不超过3个、自变量相对独立、模型能处理不固定监测天数、人为不决定参数。按监测点建模的6个独立模型都可以单独使用,解决了模型不通用的问题;可以实现逐日诊断和预测,方便与遥感信息、作物长势信息等实时匹配;差减统计模型和间隔天数统计模型精度最高,前者是基于质量守恒定律的统计模型,后者有效地解决了间隔天数不固定所带来的预测误差。     

湖泊的水生态模型

由于城市化进程的不断加快,大量的污染物排放到湖泊中,打破了湖泊的生态平衡.湖泊的水生态模型不仅能够研究湖泊的富营养化现象,而且还可以模拟工业毒物等污染物对湖泊生态系统的影响,是湖泊管理的有力工具.本文从分析湖泊的污染现状以及污染趋势出发,介绍了湖泊水生态模型的研究现状,并根据模型的模拟对象和空间模拟能力对湖泊水生态模型进行了分类.在此基础上介绍了AQUATOX、PAMOLARE、CAEDYM、WASP、OOMAS等比较成熟的模拟湖泊生态系统的模型软件,阐述了它们的产生、发展、主要特点以及应用,总结了当今湖泊水生态模型在模拟过程中遇到的问题,并对水生态模型的发展趋势进行了展望.本文为湖泊水生态模型的选择提供了建议.     

人工湿地去污模型的统一结构特征

通过对典型人工湿地去污模型的分析比较,提出了基于各模型微分方程而建立的统一去污模型。该模型能够将典型去污模型作为特例而导出,并能很好地解释这些模型之间的过渡关系。以潜流湿地中NH4^＋和BOD5的降解为例,对统一去污模型的应用情况进行了简单探讨,表明利用统一去污模型结构有助于深入揭示去污机理,从而提出更为精确的去污模型。     

景观空间动态模型研究现状和应重点解决的问题

分析了景观空间动态模型研究的现状及今后的重点发展方向。论文重点介绍了随机景观模型、邻域规则模型和景观过程模型(包括渗透模型、个体行为模型和空间生态系统模型)3类景观空间模型的发展现状、存在的主要问题以及对这些模型途径进行完善的有关工作,并从确证性分析、有效性分析与敏感性分析等3个方面阐述了模型检验技术的发展现状。最后,总结了未来景观空间动态模型发展中应重点解决的主要问题,即模型算法的优化、尺度转换、模型的复杂化与简化、模型检验与评价,通用模型的建立以及传统模型与社会经济因素衔接问题。     

畜群发展系统仿真模型的建立——黑龙江省奶牛业发展方案的研究

笔者在Lewis—Leslie结模型构的基础上,结合系统动力学思想方法,提出了分块矩阵模型。该模型除了可以研究种群结构外,还可以挂接各种投入产出信息,把生态结构模型发展成生态经济模型。本文用此模型对黑龙江省奶牛业的发展做系统仿真,提出了不同阶段的方案。     

植物种分布的模拟研究进展

从植物种水平研究植被与气候的关系一直是生态学的热点之一。该文综述了植物种与气候关系的早期研究历史和国内外近期研究进展,尤其是20世纪80年代以来,随着全球变化研究的不断发展和深入,植物种地理分布与气候因子关系研究的最新发展,汇总了最近20年来国际上模拟预测植物种潜在地理分布的模型,比较了不同模型的优缺点。统计模型主要包括以生物气候分室模型或气候分室模型为代表的相关模型、以广义线性模型和广义加性模型为代表的回归模型、以分类和回归树分析及人工神经网络为代表的基于规则的模型、以及生态位模型、气候响应面模型等。机理模型主要介绍了基于BIOME1生物地理模型和FORSKA林窗模型的STASH模型、基于过程的物候模型PHENOFIT,以及一种基于水分平衡、温度和植物物候现象的模型。总结不同模型模拟预测的不同地区植物种未来分布的格局,并介绍中国植物种潜在分布区及未来变化的模拟预测工作,从而为更加准确地模拟预测植物种在未来全球变化情景下的变化趋势提供背景知识。     

选择方向强化学习的神经网络模型

提出了神经网络模型的一种选择方向强化学习规则,定义并导出了新模型与Hopfield模型两种不同的筛选曲线,由此表明新模型对相关图样的分辨力优于Hopfield模型。在微机上模拟了由100个神经元构成的网络,结果显示新模型具有重复记忆这一神经生理学特点。定义并分行了记忆强度因子,模拟结果表明记忆强度因子愈大的记忆态,联想性能愈好,学习周期愈短。     

不同图样间联想的神经网络模型

本文讨论了Hopfield模型和其恢复特性。提出了不同图样间联想模型,推导了主态和辅态为初态时该模型的演化结果。对于30个随机图样、400个神经元组成的神经网络,给出了计算机模拟计算结果。并且讨论了不同情况下Hopfield模型和图样间联想模型的计算结果。     

地形校正对森林生物量遥感估测的影响

基于常用的4种地形校正模型（Cosine模型、C模型、C+SCS模型、Minnaert模型）,以IDL语言为二次开发平台,对黑龙江省帽儿山地区2007年7月21日TM图像进行地形校正,从视觉差异、图像的定量统计特征两方面评价了4种地形校正模型的修正效果,并比较了地形校正后几种遥感因子与森林生物量的相关性,建立了森林生物量的遥感反演模型,分析了不同地形校正模型对森林生物量反演的影响.结果表明：由于K-T变换采用线性变换方式,地形校正后遥感数据与森林生物量的相关性出现了较大波动,应根据地表信息调整变换参数,因此该变换方式不适合与地形校正结合使用;植被指数的信息量在地形校正后明显提高,其与森林生物量的相关性显著增强;4种地形校正模型中,Cosine校正过度,不宜采用,C模型和C+SCS模型通过引入半经验参数,较好地消除了地形效应,Minnaert模型校正后降低了森林生物量估测的误差,有效地提高了遥感反演模型的精度.     

Coexistence of two microbial populations competing for a renewable resource in a non-predator-prey system

If two microbial populations compete for a single resource in a homogeneous environment with time invariant inputs they cannot coexist indefinitely if the resource competed for is not renewed by biological activity within the system. Mathematical studies have shown that in a predator-prey system, where the resource (prey) is self-renewing, the two competitors (predators) can coexist in a limit cycle. This suggests that if the resource competed for is renewed by biological activity within the system coexistence can occur in any microbial system provided that it exhibits the same features as, but without being, a predator-prey one. A food chain involving commensalism, competition and amensalism is presented here. Two subcases are considered. It is only when maintenance effects are taken into account that coexistence, in limit cycles, can occur for this system. Limit cycle solutions for the system are demonstrated with the help of computer simulations. Some necessary conditions for coexistence are presented, as are some speculations regarding the possible physical explanations of the results.     

Characteristics and detection principles of a new enzyme label producing a long-term chemiluminescent signal

A new enzyme label system is described which is superior to all existing chemiluminescence labels used in immunoassays. The system consists of the enzyme xanthine oxidase with hypoxanthine as substrate. The signal reagent contains perborate, an Fe–EDTA complex and luminol. The enzyme preparation and the signal reagent are very stable upon storage. The main features of the system are a long duration of the chemiluminescent signal (half-life time of 30 hours) and a very low limit of detection (about 3 amol). Possibilities and implications for the use of various measuring system are discussed.     

Permanence of a stage-structured predator-prey system with a class of functional responses

A stage-structured predator-prey system incorporating a class of functional responses is presented in this article. By analyzing the system and using the standard comparison theorem, the sufficient conditions are derived for permanence of the system and non-permanence of predators.     

Development of a Computer-Simulation Model for a Plant-Nematode System

A computer-simulation model (MELSIM) of a Meloidogyne-grapevine system is developed. The objective is to attempt a holistic approach to the study of nematode population dynamics by using experimental data from controlled environmental conditions. A simulator with predictive ability would be useful in considering pest management alternatives and in teaching. Rates of flow and interaction between the components of the system are governed by environmental conditions. Equations for these rates are determined by fitting curves to data from controlled environment studies. Development of the model and trial simulations have revealed deficiencies in understanding of the system and identified areas where further research is necessary.     

Integration of a barcode reader with a commercial flow cytometer.

This report describes the application and installation of a barcode reader on a standard EPICS Elite flow cytometer. The barcode reader system eliminates keyboard entry of sample information on the cytometer. The system automates the transfer of sample information already present in our laboratory database to the cytometer at run time. The system uses a standard "off-the-shelf" bar code wand with a personal computer keyboard interface and requires no additional software at run time. No typing of sample information is required by the operator at any stage of normal sample operation at the cytometer. All operations are automatically coded into the cytometry software using the macro functions of the software. Tubes are inserted into the tube reader and sample information is transferred automatically into the cytometer. We have found that the system allows rapid and continuous operation of routine clinical and research samples. This automated data entry also reduces the possibility of data input errors.     

A stochstic model for a two-compartment reversible system

This paper discusses a general stochastic model for a two-compartment reversible system with non-homogeneous Poisson inputs, arbitrary residence times at each of the compartments and time-dependent transition probabilities. The probability distributions of the number of particles in each compartment and in the system are obtained together with the number of particles which depart from the system. In addition, various covariance functions with a time lag are obtained. Some of the above obtained results are deduced for time-independent arrivals, exponential residence times and time-independent transition probabilities. Fluctuations of the particles present in the system are also analysed. Similar analysis is provided for the model into which some particles are initially introduced at the system. Some possible applications are discussed at the end.     

On a stochastic model of a molecular genetic system capable of differentiation and reproduction of the initial state

A model of the system, consisting of a cyclic 3-operonic subsystem, related to a 2-operonic trigger by direct and feedback connections, is built. The operons of this system are functionally connected by repressers. The basic postulates are: a) the repressers are stable; b) during cell division each molecule of represser and respective mRNA has equal probability to find itself in any daughter cell. Computer simulation demonstrates that after few cell divisions, analogues of somatic cell, maintaining the system for some successive divisions, and of generative cells, where by the system returns to the initial state, appear. A scheme of molecular genetic control system, producing a lot of different cell types in course of successive cell divisions, is proposed.     

Steady flow of a viscous fluid through a network of tubes with applications to the human arterial system

The paper presents a finite-element model for the analysis of steady flow of a viscous fluid through a connected system of elastic tubes with the aim of simulating the conditions of blood flow through the human arterial system. The governing equations of the model are non-linear in character and are solved through an iterative computational procedure. This model is capable of incorporating the effects of stenosis on flow and pressure. Typical results are presented and discussed. Quantitative results have been obtained on blood flow through a model of the human arterial system corresponding to the sets of prescribed conditions at the terminations. Also computational results on the effect of stenosis in typical arteries of the system are presented.     

一类污染环境中非自治捕食食饵系统的持续生存与灭绝

研究了一类脉冲排放污染环境中受到人类外界干扰作用的非自治捕食食饵系统的动力学性质,利用脉冲微分方程相关理论和分析的方法给出所提出模型的持续生存与灭绝的充分条件.     

Computer control of pH and DO in a laboratory fermenter using a neural network technique

In this contribution, the advantages of the artificial neural network approach to the identification and control of a laboratory-scale biochemical reactor are demonstrated. It is very important to be able to maintain the levels of two process variables, pH and dissolved oxygen (DO) concentration, over the course of fermentation in biosystems control. A PC-supported, fully automated, multi-task control system has been designed and built by the authors. Forward and inverse neural process models are used to identify and control both the pH and the DO concentration in a fermenter containing a Saccharomyces cerevisiae based-culture. The models are trained off-line, using a modified back-propagation algorithm based on conjugate gradients. The inverse neural controller is augmented by a new adaptive term that results in a system with robust performance. Experimental results have confirmed that the regulatory and tracking performances of the control system proposed are good.