Surface soil moisture estimation from the synergistic use of the (multi-incidence and multi-resolution) active microwave ERS Wind Scatterometer and SAR data

This paper presents an original methodology to retrieve surface (<5 cm) soil moisture over low vegetated regions using the two active microwave instruments of ERS satellites. The developed algorithm takes advantage of the multi-angular configuration and high temporal resolution of the Wind Scatterometer (WSC) combined with the SAR high spatial resolution. As a result, a mixed target model is proposed. The WSC backscattered signal may be represented as a combination of the vegetation and bare soil contributions weighted by their respective fractional covers. Over our temperate regions and time periods of interest, the vegetation signal is assumed to be principally due to forests backscattered signal. Then, thanks to the high spatial resolution of the SAR instrument, the forest contribution may be quantified from the analysis of the SAR image, and then removed from the total WSC signal in order to estimate the soil contribution. Finally, the Integral Equation Model (IEM, [IEEE Transactions on Geoscience and Remote Sensing, 30 (2), (1992) 356]) is used to estimate the effect of surface roughness and to retrieve surface soil moisture from the WSC multi-angular measurements. This methodology has been developed and applied on ERS data acquired over three different Seine river watersheds in France, and for a 3-year time period. The soil moisture estimations are compared with in situ ground measurements. High correlations (R² greater than 0.8) are observed for the three study watersheds with a root mean square (rms) error smaller than 4%.     

多策略多参数并行差分进化算法

针对差分进化（DE）算法收敛早熟与计算效率不理想的问题,提出一种改进的差分进化算法。首先,在进化中同时并行多个策略与参数组合来提高个体多样性。其次,依据建立的评价指标自适应地调整组合来提高寻优效率。最后,把进化过程分为若干的子进程以避免前期优势组合不适应后期的问题。在10个标准测试函数上的实验结果表明,提出的算法与其他算法相比具有相对较好的性能。     

动态聚类适应值共享遗传算法在散射计风矢量反演中的应用

风矢量反演是散射计数据处理的核心内容,传统风矢量反演算法的设计过多依赖于目标函数的具体分布形态。以SeaWinds散射计为例,根据风矢量反演的多解问题和模糊解特性,设计了一种基于动态小生境遗传算法的风矢量反演算法。利用部分L2A和相应L2B数据对该算法进行了验证。结果表明该算法在无需任何目标函数先验知识的条件下能够取得较好的反演结果。     

Mapping understory vegetation using phenological characteristics derived from remotely sensed data

Understory vegetation is an important component in forest ecosystems not only because of its contributions to forest structure, function and species composition, but also due to its essential role in supporting wildlife species and ecosystem services. Therefore, understanding the spatio-temporal dynamics of understory vegetation is essential for management and conservation. Nevertheless, detailed information on the distribution of understory vegetation across large spatial extents is usually unavailable, due to the interference of overstory canopy on the remote detection of understory vegetation. While many efforts have been made to overcome this challenge, mapping understory vegetation across large spatial extents is still limited due to a lack of generality of the developed methods and limited availability of required remotely sensed data. In this study, we used understory bamboo in Wolong Nature Reserve, China as a case study to develop and test an effective and practical remote sensing approach for mapping understory vegetation. Using phenology metrics generated from a time series of Moderate Resolution Imaging Spectroradiometer data, we characterized the phenological features of forests with understory bamboo. Using maximum entropy modeling together with these phenology metrics, we successfully mapped the spatial distribution of understory bamboo (kappa: 0.59; AUC: 0.85). In addition, by incorporating elevation information we further mapped the distribution of two individual bamboo species, Bashania faberi and Fargesia robusta (kappa: 0.68 and 0.70; AUC: 0.91 and 0.92, respectively). Due to its generality, flexibility and extensibility, this approach constitutes an improvement to the remote detection of understory vegetation, making it suitable for mapping different understory species in different geographic settings. Both biodiversity conservation and wildlife habitat management may benefit from the detailed information on understory vegetation across large areas through the applications of this approach.     

含两个参数的奇异摄动问题的差分进化算法

针对在Shishkin网格上数值求解含有两个参数的奇异摄动问题,在有限差分方法的基础上,将Shishkin网格过渡点参数选取问题转化成一个无约束优化问题,并采用差分进化算法进行求解。数值结果表明用差分进化算法得到最优Shishkin网格参数后,奇异摄动问题的数值解在边界层的精度得到了明显的提高,进一步说明了方法的有效性和可靠性。     

进化粒子群优化算法及其在各向异性含水层参数估计中的应用

在粒子群算法的基础上,引入进化思想和组群思想提出一种新的智能优化算法———进化粒子群算法( EPSO)。基于抽水试验数据,将EPSO算法应用到各向异性含水层参数估计中,对算法性能进行研究并与其他算法进行了对比,发现标准粒子群算法及其一般改进算法已不能有效求解各向异性含水层参数,而EPSO算法进行多次计算后,1)结果可靠;2)目标函数值及待估参数稳定;3)对初始范围的鲁棒性好。结果表明EPSO算法对各向异性含水层参数估计问题具有可靠性、收敛性和鲁棒性,可望应用到更广泛的参数识别问题中。     

小数据集条件下贝叶斯网络自适应参数学习方法

针对小数据集条件下贝叶斯网络参数学习问题,约束最大似然(CML)和定性最大后验概率(QMAP)方法是两种约束适用性较好的方法.当样本数量、约束数量、参数位置不同时,上述两种方法互有优劣,进而导致方法上的难以选择.因此,本文提出一种自适应参数学习方法:首先,利用CML和QMAP方法学习得到两组参数;然后,基于拒绝–接受采样和空间最大后验概率思想自定义计算得到样本权重、约束权重、参数位置权重;最后,基于上述参数和权重计算得到新的参数解.实验表明:在任何条件下,本文方法计算得到参数的精度接近甚至优于CML和QMAP方法的最优解.     

Learning Bayesian network parameters under incomplete data with domain knowledge

Bayesian networks (BNs) have gained increasing attention in recent years. One key issue in Bayesian networks is parameter learning. When training data is incomplete or sparse or when multiple hidden nodes exist, learning parameters in Bayesian networks becomes extremely difficult. Under these circumstances, the learning algorithms are required to operate in a high-dimensional search space and they could easily get trapped among copious local maxima. This paper presents a learning algorithm to incorporate domain knowledge into the learning to regularize the otherwise ill-posed problem, to limit the search space, and to avoid local optima. Unlike the conventional approaches that typically exploit the quantitative domain knowledge such as prior probability distribution, our method systematically incorporates qualitative constraints on some of the parameters into the learning process. Specifically, the problem is formulated as a constrained optimization problem, where an objective function is defined as a combination of the likelihood function and penalty functions constructed from the qualitative domain knowledge. Then, a gradient-descent procedure is systematically integrated with the E-step and M-step of the EM algorithm, to estimate the parameters iteratively until it converges. The experiments with both synthetic data and real data for facial action recognition show our algorithm improves the accuracy of the learned BN parameters significantly over the conventional EM algorithm.     

Adaptive differential evolution algorithm with novel mutation strategies in multiple sub-populations

Differential evolution (DE) algorithm has been shown to be a very effective and efficient approach for solving global numerical optimization problems, which attracts a great attention of scientific researchers. Generally, most of DE algorithms only evolve one population by using certain kind of DE operators. However, as observed in nature, the working efficiency can be improved by using the concept of work specialization, in which the entire group should be divided into several sub-groups that are responsible for different tasks according to their capabilities. Inspired by this phenomenon, a novel adaptive multiple sub-populations based DE algorithm is designed in this paper, named MPADE, in which the parent population is split into three sub-populations based on the fitness values and then three novel DE strategies are respectively performed to take on the responsibility for either exploitation or exploration. Furthermore, a simple yet effective adaptive approach is designed for parameter adjustment in the three DE strategies and a replacement strategy is put forward to fully exploit the useful information from the trial vectors and target vectors, which enhance the optimization performance. In order to validate the effectiveness of MPADE, it is tested on 55 benchmark functions and 15 real world problems. When compared with other DE variants, MPADE performs better in most of benchmark problems and real-world problems. Moreover, the impacts of the MPADE components and their parameter sensitivity are also analyzed experimentally.     

通用遗传算法估算化学及生化反应动力学参数

遗传算法是一种借鉴生物界自然选择和进化机制发展起来的高度并行、随机、自适应的搜索方法。MATLAB通用遗传算法工具箱GAOT本着使用群体搜索技术,将种群代表一组问题的解,通过对当前种群施加选择、交叉和变异等一系列遗传操作,从而得到新的一代种群,并逐步使种群进化到包含近似最优解状态的原则,对苯热裂解脱氢反应及产赖氨酸分批发酵动力学模型参数进行了估算。与文献中已有的结果相比,模型计算值与实验值的吻合程度相似或更优。所用方法也可用于以微分方程组为数学模型的这类过程的参数估算或寻优,从而研究者可以更加集中注意力于深入解决与数学模型本身有关的问题。     

一种增量更新算法在数据挖掘中的应用

引入扩展差别矩阵和扩展决策矩阵,提出了新的属性约简算法和增量更新算法,即基于扩展差别矩阵的属性约简算法和基于扩展决策矩阵的增量式规则提取算法,讨论了规则的增量更新算法。由于使用了增量更新算法和并行处理技术,从而提高了数据挖掘的效率,降低了时间复杂度。通过实验说明此算法是有效和可行的。     

加权变异策略动态差分进化算法

针对差分进化算法在解决高维优化问题时易早熟收敛、求解精度低和参数设置麻烦等问题,提出一种加权变异策略动态差分进化算法（WMDDE）。为了动态平衡全局搜索与局部搜索能力,跳出局部最优,将标准差分进化算法的变异策略DE/rand/1和DE/best/1进行加权组合,提出两种新的随机扰动加权变异算子。提出一种动态自适应调整缩放因子和交叉概率因子的策略,避免参数设置的麻烦,提高算法的稳定性。在11个Benchmark函数上的测试结果表明,新算法能有效避免早熟收敛,全局寻优能力强,且在高维时寻优速度、求解精度和稳定性均优于4种DE进化算法。     

高通信效率的分布式流数据聚类算法

传感器节点的资源是有限的,高的通信开销会消耗大量的电量。为了减小分布式流数据分类算法的通信开销,提出一种高效的分布式流数据聚类算法。该算法包含在线局部聚类和离线全局协同聚类两个阶段。在线局部聚类算法将每个流数据源进行局部聚类,并将聚类后的结果通过序列化技术发往协同节点;协同节点得到来自不同流数据源的局部聚类信息后进行全局聚类。从实验中可以看出,当不断增加窗口的大小时,算法用于数据发送的时间恒定不变,算法的聚类时间和总的时间呈线性增长,即所提出算法的执行时间不受滑动窗口宽度和聚类个数的影响;同时该算法与集中式算法的准确性接近,并且通信开销远远小于相关的分布式算法。实验结果表明,该算法具有很好的可扩展性,可应用于对大规模分布式流数据源进行聚类分析。     

A differential evolution algorithm with self-adapting strategy and control parameters

This paper presents a Differential Evolution algorithm with self-adaptive trial vector generation strategy and control parameters (SspDE) for global numerical optimization over continuous space. In the SspDE algorithm, each target individual has an associated strategy list (SL), a mutation scaling factor F list (FL), and a crossover rate CR list (CRL). During the evolution, a trial individual is generated by using a strategy, F, and CR taken from the lists associated with the target vector. If the obtained trial individual is better than the target vector, the used strategy, F, and CR will enter a winning strategy list (wSL), a winning F list (wFL), and a winning CR list (wCRL), respectively. After a given number of iterations, the FL, CRL or SL will be refilled at a high probability by selecting elements from wFL, wCRL and wSL or randomly generated values. In this way, both the trial vector generation strategy and its associated parameters can be gradually self-adapted to match different phases of evolution by learning from their previous successful experience. Extensive computational simulations and comparisons are carried out by employing a set of 19 benchmark problems from the literature. The computational results show that overall the SspDE algorithm performs better than the state-of-the-art differential evolution variants.     

求解药代动力学参数的混合人口迁移算法

针对传统方法优化药代动力学参数时精度不高的缺陷,将Hooke-Jeeves算法与人口迁移算法有机融合,使两者取长补短,既提高了算法的精度,又加快了算法的收敛速度。将混合人口迁移算法用于血管外给药二室模型参数优化的实验之中,不仅比传统的残数法效果要好,而且比Hooke-Jeeves算法或人口迁移算法更优,精度更高。多次实验表明:算法具有良好的可靠性和稳定性,是一种较好的解决药代动力学参数的方法。     

A novel hybrid multi-objective immune algorithm with adaptive differential evolution

In this paper, we propose a novel hybrid multi-objective immune algorithm with adaptive differential evolution, named ADE-MOIA, in which the introduction of differential evolution (DE) into multi-objective immune algorithm (MOIA) combines their respective advantages and thus enhances the robustness to solve various kinds of MOPs. In ADE-MOIA, in order to effectively cooperate DE with MOIA, we present a novel adaptive DE operator, which includes a suitable parent selection strategy and a novel adaptive parameter control approach. When performing DE operation, two parents are respectively picked from the current evolved and dominated population in order to provide a correct evolutionary direction. Moreover, based on the evolutionary progress and the success rate of offspring, the crossover rate and scaling factor in DE operator are adaptively varied for each individual. The proposed adaptive DE operator is able to improve both of the convergence speed and population diversity, which are validated by the experimental studies. When comparing ADE-MOIA with several nature-inspired heuristic algorithms, such as NSGA-II, SPEA2, AbYSS, MOEA/D-DE, MIMO and D²MOPSO, simulations show that ADE-MOIA performs better on most of 21 well-known benchmark problems.     

基于不完全数据的数据挖掘方法研究

将不完全数据分为了两类：属性值残缺和属性值隐含.对基于这两类不完全数据的数据挖掘方法分别进行了探讨,给出了相应的处理方法,并对这些方法及其应用进行了讨论.属性值残缺的处理主要采用一系列＂补漏＂的方法,使数据成为完全数据集;属性值隐含的处理则通过EM算法来优化模型的参数,弥补数据的不完全性.     

C-correction of optical satellite data over alpine vegetation areas: A comparison of sampling strategies for determining the empirical c-parameter

Semi-empirical topographic normalization methods (e.g., C-correction) have been widely used to correct illumination differences in optical satellite data. The objective of this study was to examine the precision and accuracy of the C-correction's empirical parameter, c, as a function of the sample from which it was derived. Three sampling methods were compared: a random sample, a sample stratified on north and south aspects, and a sample stratified by cosine of the solar incidence angle, i. In the latter, power allocation was used to determine the quantity of observations for each stratum. Four overlapping satellite images were used (two Landsat 5 TM and two SPOT 5 HRG) with different acquisition dates and large solar zenith angles over an alpine region in Sweden. The sample stratified by cosine of i produced c with the highest precision from repeated trials and had coefficients of determination (R²) twice as high as those from the other sampling methods. Use of power allocation in the cosine of i stratified sample enabled better representation of spectral variability; this was particularly important for the NIR band where the outcome of c differed according to sampling method. Evaluations using t-tests and classification accuracy showed that c derived from the cosine of i stratified sample correctly normalized a larger percentage of the evaluation data. The distribution of cosine of i in the study area, the spectral variability and vegetation types exert influences to consider when sampling to derive c. Although sampling was restricted to alpine vegetation only, some vegetation classes may have benefitted from separate c-parameter calculation. In general, dry alpine heath and alpine grass heath had relatively higher c-parameters, mesic alpine heath was slightly lower, and alpine willow and alpine meadow had lower c-parameters for the near-infrared band. The cosine of i stratified sampling method using power allocation may be useful for calculation of c for vegetation conditions other than those presented here, as well as for other empirical parameters (e.g., the Minnaert constant, k).     

Box Car过程数据压缩算法性能分析

Box Car过程数据压缩算法在现场总线控制系统中得到广泛采用。其压缩效果受记录限和压缩区间的影响。本文基于对典型仿真数据的大量计算,分析了Box Car过程数据压缩算法记录限和压缩区间对趋势平稳的过程数据的压缩比、计算时间和压缩系数的影响。本文还分析了过程数据趋势特征和波动特性对Box Car算法压缩比和逼近系数的影响。本文的计算结果对于在实际应用中根据过程数据不同的趋势和噪声特征调整Box Car压缩算法参数以获得理想的压缩效果具有指导意义。     

一种基于改进的DBSCAN的面向海量船舶位置数据码头挖掘算法

目前我国正在大力推行"一带一路"航海战略,航海事业蓬勃发展,大量新码头正在修建中。如何快速、准确地更新码头的空间信息,对于分析进出口贸易、提高码头服务效率等具有很强的现实意义。当前我国主要通过人工测绘手段更新海图,更新间隔在3~12月,远不能满足需求。而利用包括国际海事卫星C系统、北斗卫星、Argos卫星等手段获取的船舶位置数据来进行码头挖掘,为解决获得码头空间信息问题提供了新手段。利用自动识别系统AIS获取的海量船舶位置数据,提出了一种基于自优化参数的码头挖掘算法DBSCAN。一方面能够面向不同船舶类型的不同密度分布进行自动学习优化DBSCAN核心参数,进而聚类出包含码头的停泊区域,具备很强的灵活性;另一方面,融合岸基结构物等空间数据,对停泊区域中的锚区和临时停泊区域等进行排除,获取码头的空间信息,并且达到很高的准确率。利用2012年4月至2014年4月两年中国滚装船的真实轨迹数据和国际滚装船真实轨迹数据进行了码头挖掘实验,准确率能够达到93%以上。