基于种群聚类和互信息的入侵检测方法

网络中爆炸式增长的数据对于数据分析处理方法提出了更大的挑战,尤其在入侵检测领域。针对网络数据样本量的持续增大和数据维度的迅速增加问题,首先从特征选择的角度入手,提出一种基于种群聚类的特征选择方法,定义相似度函数和优度值函数,实现粒子群算法初始种群的均匀分布;其次,将基于互信息的信息度量方法集成到基于种群聚类的特征选择方法中,进一步从局部与全局相结合优化的角度,实现所选特征与目标类的最大相关;最后采用支持向量机作为分类器,并定义包含检测率和误报率的双目标函数作为适应度函数,进行入侵检测。实验结果表明,该算法可以有效地避免数据过早收敛,且提高了入侵检测的准确率和检测率。

基于聚集密度的约束多目标进化算法

对基于群体聚类的约束多目标进化算法进行了改进,引入了聚集密度以度量群体中个体间的关系,保持种群的多样性。其基本思想为:首先将初始群体按多判据聚类方法分为适应度值不同的四类,然后计算类内群体中个体的聚集密度,根据适应度值和聚集密度定义一个偏序集,最后采用比例选择原则依次从偏序集中选择个体,更新精英集。通过数值实验用量化指标研究了改进算法的收敛性和分布性,结果表明:改进算法的收敛性与常规约束多目标进化算法相当,但分布性有了明显的提高。

基于改进多目标粒子群优化算法的长距离城乡输水管道优化

长距离输水管道在水资源配置中具有重要作用,其优化设计直接影响到水资源的输送效率。本文针对长距离城乡供水管道系统的优化问题,提出了一种基于改进多目标粒子群优化算法(MOPSO)的新方法。综合优化输水管段直径和输水线路,以费用投资最小和流量保证率最大为目标函。结果表明,所提方法相较于传统粒子群优化算法流量保证率提升了2.25%,管道费用减小了225.64万元,并且在多个性能指标上显著优于目前经典的多目标优化方法。研究结果为长距离输水管道设计提供参考。

Cluster Analysis of Morphologic Characteristic of Eight Geographical Populations of Rana Dybowskii

[ObJective] The research aimed to determine the geographic distribution map of system of Rana dybowskii. [Method] Four morphologic indices （body length, body weight, forelimb length, hindlimb length） of eight geographical populations of R.dybowskii which naturally distribute in Changhai Mountain and Xiaoxing＇an Mountain were measured. Measure results were variance analyzed and cluster analyzed. [Result] Variance analysis showed： the genetic branching among the Dongfanghong male population（ belongs to Wandashan） and Xiaoxing＇an Mountain male population and Changbai Mountain male population were significantly different （P〈0.05） ; the genetic branching between the Hebei female population （belongs to Xiaoxing＇an Mountain） and Changbai Mountain female population was significantly different （P〈0.05 ）. Cluster analysis showed ： male R.dybowskii can be divided into three groups ： the first group included Quanyang, Tianbei, Chaoyang and Ddkouqin, the second group included Tieli and Anshan, the third group included Dongfanghong; and the female R. dybowskii can be divided into three groups ： the first group included Quanyang and Chaoyang, the second group included Tianbei and Dakouqin, the third group included Hebei. [Condusion] The paper deduced that the Sanjiang Plain was the geographical origin center ofR. dybowskii which radiated to Changbai Mountain and Xiaoxing＇an Mountain along the adverse current of Songhua River basin, therefore, the current distribution pattern of R. dybowskii was formed.

Spatial and temporal variation of fish assemblages in a subtropical small stream of the Huangshan Mountain

Spatial and temporal variation of fish assemblages were investigated seasonally from May 2007 to February 2008 across 11 study sites in a subtropical small stream, the Puxi Stream, of the Huangshan Mountain. Along the longitudinal gradient from headwater to downstream, fish species richness and abundance increased gradually, but then decreased significantly at the lower reaches. The highest species richness and abundance were observed in August and the lowest in February. Based on analysis of similarities （ANOSIM）, fish assemblages were significantly different in spatial variation but not in temporal variation. Although differences were observed both among sites and among stream orders, the lower R value in order-variation suggested stream order was not the optimal factor explaining the spatial variation of fish assemblages. In addition, dam construction did not significantly alter fish assemblages in the sites adjacent to and immediately downstream to dams. Using cluster analysis and non-metric Multi Dimensional Scaling analysis （NMS）, assemblages were separated into three groups at a Bray-Curtis similarity value of 42%： the upper, middle and lower groups. Following analysis of similarity percentages of species contributions （SIM- PER）, shifts in occurrence or abundance of S. curriculus, Z. platypus, R. bitterling and A. fasciatus contributed most to the differences amongst the three groups. Standard Deviation Redundancy Analysis （RDA） suggested that habitat structure （such as elevation, substrate, and flow velocity） contributed to the spatial and temporal pattem of fish assemblages in the Puxi Stream. In conclusion, the fish assemblages in Puxi Stream presented significant spatial but not temporal variation. Human disturbance has perhaps induced the decrease in species diversity in the lower reaches. However, no significant change was observed for fish assemblages in sites far from and immediately downstream from low-head dams [Current Zoology 56 （6）： 670-677, 2010].

Stability of Mbuna Species Populations in Lake Malawi

The first comprehensive study of abundance, richness and distribution of Mbuna in Lake Malawi was carried out in the 1980s. The present study examined changes of species richness and abundance of Mbuna with time. Most sites demarcated during the 1980s survey were explored in order to determine changes in abundance and composition of Mbuna over a period of more than 20 years Chi-square analysis revealed significant difference （p 〈 0.05） in abundance of Mbuna between 1983 and 2005 with the exception of Chinyankhwazi and Chinyamwezi where the difference was not significant （p 〉 0.05）. Cluster analysis showed changes in species composition between 1983 and 2005 in various sites. Generally, there were more species lost in a given area than new species observed in 2005. Decline in Mbuna species richness might have been caused by siltation, harvesting of Mbuna for food and ornamental fish trade while new species were mostly introduced through translocation from one part of the lake to another.

Influences of sea ice on eastern Bering Sea phytoplankton

The influence of sea ice on the species composition and cell density of phytoplankton was investigated in the eastern Bering Sea in spring 2008. Diatoms, particularly pennate diatoms, dominated the phytoplankton community. The dominant species were Grammonema islandica （Grunow in Van Heurck） Hasle, Fragilariopsis cylindrus （Grunow） Krieger, F. oceanica （Cleve） Hasle, Navicula vanhoeffenii Gran, Thalassiosira antarctica Comber, T. gravida Cleve, T. nordenskioeldii Cleve, and T. rotula Meunier. Phytoplankton cell densities varied from 0.08× 10^4 to 428.8× 10^4 cells/L, with an average of 30.3× 10^4 cells/L. Using cluster analysis, phytoplankton were grouped into three assemblages defined by ice-forming conditions： open wate.r, ice edge, and sea ice assemblages. In spring, when the sea ice melts, the phytoplankton dispersed from the sea ice to the ice edge and even into open waters. Thus, these phytoplankton in the sea ice may serve as a “seed bank” for phytoplankton population succession in the subarctic ecosystem. Moreover, historical studies combined with these results suggest that the sizes of diatom species have become smaller, shifting from microplankton to nannoplankton-dominated communities.