基于神经网络的活性污泥系统建模及其控制

针对松江污水厂污水处理活性污泥系统,采用神经网络技术进行建模试验研究,在对实际运行数据剔除异常数据后,将样本数据随机分成训练样本、检验样本和测试样本.用试凑法确定合理的神经网络隐层节点数,用检验样本实时监控训练过程从而避免"过训练"现象,用多次改变网络初始连接权值求得全局极小点,从而建立了泛化能力较好的基于神经网络的活性污泥系统数学模型.利用建立的神经网络模型,对活性污泥系统运行情况的仿真与控制进行了分析研究.示例研究表明:神经网络技术能较好地应用于活性污泥系统的建模与控制,有很好的理论与实践意义.     

Modeling Prescribed Surface-Fire Regimes for Pinus strobus Conservation

Abstract:  We developed a simple model of   Pinus strobus L. stand dynamics to compare the impacts of different temporal arrangements of surface fires designed to reflect the application of fire as both an essential ecosystem process (natural fire) and as an efficient means of producing specific habitat features or other values (optimal fire). We used a stochastic simulation model of fire processes to estimate the mean fire-return interval that would maximize stand structural diversity. We investigated trade-offs between structural diversity and temporal population stability associated with changes in the fire interval and used a deterministic version of the model to explore the effects of scheduling fires at fixed intervals. In stochastic simulations, maximum structural diversity occurred at intermediate levels of disturbance (40-year mean fire interval). When fires were scheduled at fixed intervals, a longer, 100-year return interval maximized diversity. Mean fire-return interval was a mitigating factor in the diversity-stability relationship, which changed from positive to negative as the fire interval was reduced progressively from 250 to 5 years. As an alternative to scheduling fires at specified mean intervals, we developed a goal-programming model (a form of linear programming model) and used it to identify an optimal fire schedule for achieving habitat and visual-quality objectives. In comparison with the 40-year stochastic mean fire interval, which maximized structural diversity, the optimal schedule produced comparable levels of both diversity and fire frequency. Our results show how simulation and goal-programming models can be used to evaluate prescribed fire-scheduling alternatives and to explore the comparative advantages of natural and optimal fire-management approaches.     

A dynamic model of mating behavior in digger wasps: the energetics of male-male competition mimic size-dependent thermal constraints

I develop a state-based dynamic model of behavior to demonstrate that size-dependent differences in temperature tolerances are not necessary to account for the activity of small male digger wasps late in the day. In the model, males defend or patrol the nesting area, wait near nests, or feed away from the nesting area depending on time of day, energy reserves and size rank. I assume a large male competitive advantage, so mating opportunities decrease with size rank for territorial or patrolling males and are rare for all waiting males; the costs of patrolling or defense are higher than the costs of waiting. If energy reserves of all males are initially small, all males alternate feeding and territorial or patrolling behavior. If energy reserves are initially large, large males patrol or maintain territories until they risk starvation and leave the area to feed. At this time, smaller males that have conserved their resources by waiting and feeding may defend territories or patrol. I simulate the behavior of three populations representing two species of Microbembex by assuming large initial energy reserves for populations in which males were territorial and small initial reserves for populations in which males patrolled, and then convert the predicted time of activity to temperature using local regressions from field studies. Temporal patterns in the activity of large and small males were similar to those actually observed, and relationships between size and temperature predicted by the model corresponded to most observations and were sometimes positive. Thus, the delayed activity of smaller males does not correspond to activity at higher temperatures and is probably not attributable to size-dependent thermal tolerances, but may represent a temporal displacement of mating activity due to intra-sexual competition and mediated by energetics. The model makes testable predictions on the timing of feeding and depletion of energy reserves in relation to size and initial energy state, and suggests how differences among species may influence the temporal and spatial organization of male mating behavior. Received: 27 February 1997 / Accepted after revision: 26 July 1997     

Wind Tunnel Experiments and Numerical Simulation of Snow Drifting around an Avalanche Protecting Dam

To learn about wind flow and snow drifting around avalanche dams, experiments were done in the Jules Verne Climatic Wind Tunnel. The paper reports the results from numerical wind flow simulations that were done to support the findings from the wind tunnel. Satisfying the model similitude criteria for the wind tunnel configuration was difficult due to the inevitable small geometric scale of the model, while on the contrary the snow drifting conditions in the facilities were full scale. By comparing numerical wind flow results of full scale and model scale dams with the snow pattern observed in the wind tunnel, it was possible to conclude that albeit poor model similitude, the snowdrifts on the windward side of the wind tunnel model are likely to indicate the full scale natural situation.     

The tactics of mutual mate choice and competitive search

A model of mutual mate choice is described, formulated as a dynamic game, which yields predictions about mating behaviour under the influence of time constraints, choice costs and competition for mates. These variables were examined because they may result in a change in the distribution of qualities among unmated individuals of both sexes over the course of the breeding season. The model predicts that mutual choice gives rise to assortative mating, although high costs of choice and/or inaccurate assessment both lead to lower overall correlations between the qualities (or the attractiveness) of mates. When all individuals are present from the start of the breeding season, the correlation between the qualities of individuals pairing at a given time declines throughout the season, so that mates are more closely matched among individuals who pair early than among those who pair late (and extra-pair copulation may thus be more common among the latter). Delayed arrival of lower-quality individuals may, however, lead to an increase in this correlation with time during the early part of the season. The mean quality of unmated males and females declines over time, because more attractive individuals tend to mate sooner. As a result of this decline, and because of time constraints, superior individuals become less choosy as the season progresses. If choice is costly, however, then inferior individuals become more selective with time during the early part of the season, and the level of choosiness peaks later for such individuals. Received: 28 March 1996 / Accepted after revision: 21 September 1996     

Efficiency of composite sampling for estimating a lognormal distribution

In many environmental studies measuring the amount of a contaminant in a sampling unit is expensive. In such cases, composite sampling is often used to reduce data collection cost. However, composite sampling is known to be beneficial for estimating the mean of a population, but not necessarily for estimating the variance or other parameters. As some applications, for example, Monte Carlo risk assessment, require an estimate of the entire distribution, and as the lognormal model is commonly used in environmental risk assessment, in this paper we investigate efficiency of composite sampling for estimating a lognormal distribution. In particular, we examine the magnitude of savings in the number of measurements over simple random sampling, and the nature of its dependence on composite size and the parameters of the distribution utilizing simulation and asymptotic calculations.     

Renormalized numerical simulation of flow over planar and non-planar fractal trees

We review the fundamentals of a new numerical modeling technique called Renormalized Numerical Simulation (RNS). The goal of RNS is to model the drag force produced by high Reynolds-number turbulent flow over objects that display scale-invariant properties, objects such as tree-like fractals. The hallmark of RNS in this application is that the drag of the unresolved tree branches is modeled using drag coefficients measured from the resolved branches and unresolved branches (as modeled in previous iterations of the procedure). In the present paper, RNS is used to study the effects of branch orientation on the drag force generated by highly idealized trees in which trunk and branches have square cross-section, and the branches all lie in a plane perpendicular to the incoming flow. Then, the procedure is generalized to the more general case of non-planar branch arrangements. Results illustrate that RNS may enable numerical modeling of environmental flow processes associated with fractal geometries using affordable computational resolution.     

Bayesian areal wombling via adjacency modeling

Recently, public health professionals and other geostatistical researchers have shown increasing interest in boundary analysis, the detection or testing of zones or boundaries that reveal sharp changes in the values of spatially oriented variables. For areal data (i.e., data which consist only of sums or averages over geopolitical regions), Lu and Carlin (Geogr Anal 37: 265–285, 2005) suggested a fully model-based framework for areal wombling using Bayesian hierarchical models with posterior summaries computed using Markov chain Monte Carlo (MCMC) methods, and showed the approach to have advantages over existing non-stochastic alternatives. In this paper, we develop Bayesian areal boundary analysis methods that estimate the spatial neighborhood structure using the value of the process in each region and other variables that indicate how similar two regions are. Boundaries may then be determined by the posterior distribution of either this estimated neighborhood structure or the regional mean response differences themselves. Our methods do require several assumptions (including an appropriate prior distribution, a normal spatial random effect distribution, and a Bernoulli distribution for a set of spatial weights), but also deliver more in terms of full posterior inference for the boundary segments (e.g., direct probability statements regarding the probability that a particular border segment is part of the boundary). We illustrate three different remedies for the computing difficulties encountered in implementing our method. We use simulation to compare among existing purely algorithmic approaches, the Lu and Carlin (2005) method, and our new adjacency modeling methods. We also illustrate more practical modeling issues (e.g., covariate selection) in the context of a breast cancer late detection data set collected at the county level in the state of Minnesota.     

Comparison of Spatial Methods for the Assessment of Planktonic Patches

Planktonic patches are defined as areas where the abundance of plankters is above a threshold value τ. The estimation of patch size and shape can be approached using spatial statistical tools, using truncated random fields or indicator random fields as classifiers. In all cases there is the risk of false positive and false negative errors. In this paper we present the results of a comparative study on the performance of four commonly used methods: conditional simulation and kriging, both in the original measurement units of the data and under an indicator transform. We used a misclassification cost function to compare the four methods. Our results show that conditional simulation in the original measurement units attains the lowest misclassification cost. We also illustrate how the point at which this minimum is attained can be used to chose an optimal cut-off value for binary classification. Received: December 2003 / Revised: June 2005     

用平板双分子层脂膜模拟研究锌离子跨细胞膜的传递过程

制备了氧化胆固醇-卵磷脂（脑磷脂）平板双分子层脂膜，研究了膜配方对双分子层脂膜的稳定性和离子传递性能的影响，得到了最佳制膜工艺，并建立了锌离子跨卵（脑）磷脂膜的吸附-扩散模型，其计算值与实验值基本吻合．