小油菜不同播期与密度下黄条跳甲为害情况的研究

本文报道了黄条跳甲类害虫在油菜不同播期与密度条件下的发生为害情况，从四个播种日期与五个播种密度的试验结果表明，黄条跳甲随播期推迟发生为害加重，播种量大，为害较严重，并且，在西宁地区为害小油菜的主要种类是黄宽条跳甲，其次为黄曲条跳甲，二者比例为３．３：１，适时早播，严格控制播种量是减轻黄条跳甲发生为害的重要农业技术措施。     

柑桔缠皮虫Agrilus inamoenus Kerremans的鉴定和生物学简记(鞘翅目:吉丁虫科)

本文鉴定描述一种柑桔吉丁虫——柑桔缠皮虫的学名、形态特征.并简要记述该虫在福建的发生情况.     

沙漠甲虫背部凝水分析

基于沙漠甲虫利用由亲水突起和疏水凹槽组成的背部凝聚空气中水分的现象，分析凝水过程物理机理，给出数学描述．将凝水过程分作三个阶段，吸聚：小雾滴在背部亲水突起上吸附，长大成小水珠；聚合长大：单突起上直径为1—2mm的小水珠因不能脱落而与邻近水珠聚合，之后长大为大水珠；脱落：大水珠从背部脱落，流入甲虫口中．导出了水珠脱落半径与甲虫背部表面及外界环境参数的数学关系，得出风速为5m／s时，直径至4．1--5．6mm的水珠以滑动方式脱落，这与文献中实际观察到的水珠脱落直径为4—5mm基本一致．     

基于MFCC与神经网络的小蠹声音种类自动鉴别

昆虫发出的各种声音具有种间特异性,是非常可靠的分类依据。利用这一特性,本实验旨在探索一种对昆虫自动分类的新方法。本实验录制了红脂大小蠹、云南切梢小蠹、短毛切梢小蠹和华山松大小蠹4种小蠹虫的胁迫声,利用Adobe Adition2.0对每个声音文件进行降噪,再将其截取成只含有一个脉冲组的声音片段。在MATLAB环境下对这些声音片段进行端点监测并提取12维的MFCC（Mel频率倒谱系数）,然后将此特征参数输入BP神经网络进行训练和检测。设置训练样本数为20、40、60、80、100,4种小蠹检测样本数分别为54、95、54、50,结果显示识别率随着训练样本数的增加而提高,在训练样本量为100时的最高识别率达到98.14%,平均识别率为93.29%,收到了较好的效果。为了验证小蠹种类数对识别率的影响,本实验对4种小蠹进行了两两比较,结果显示总体上高于4种一起识别的结果。     

重采样和天牛须协同演化粒子群的WSN覆盖控制算法

为了最大程度提升无线传感器网络(WSN)的覆盖范围并降低能耗,延长网络生命周期,提出了基于重采样技术和天牛须搜索的协同演化粒子群优化(RBASPSO)算法来优化WSN的覆盖控制问题。重采样技术平衡了粒子群算法的全局搜索能力和收敛速度,增加了粒子群整体多样性,防止算法过早收敛,加强粒子在搜索过程中跳出低质量谷底的能力; 天牛须搜索依靠个体的两个触角搜索其邻域,增强了粒子群中单个粒子的搜索能力。RBASPSO算法采用覆盖率和节点休眠率的加权作为优化WSN覆盖控制的目标函数,通过重采样技术和天牛须搜索的协同演化,既加强了单个粒子的搜索能力,又确保粒子群的多样性及活跃性,提升WSN覆盖性能。实验结果表明,RBASPSO算法不仅能有效处理复杂多峰问题; 而且可以有效提高WSN网络覆盖率,延长网络生命周期。     

OPTIMAL TIMING OF SALVAGE HARVEST IN RESPONSE TO A STOCHASTIC INFESTATION

Abstract Following a catastrophic disturbance, forest managers may choose to perform a salvage harvest to recoup timber losses. When the disturbance process evolves stochastically, a unique option value arises associated with the salvage harvest decision. This option value represents the value of postponing a salvage harvest to gain more information about the disturbance process. This paper uses a real options approach to determine how much of a forested area must be infested to trigger a salvage harvest when the forest provides both timber and nontimber values. Analytical results indicate slower rates of forest recovery will optimally delay a salvage harvest while forested areas with large timber values and where nontimber values are more sensitive to the presence of dead and dying trees should be harvested more immediately. The model is applied to a mountain pine beetle outbreak in Idaho's Sawtooth National Forest using readily available aerial detection survey data.     

Stable striped state in a rotating two-dimensional spin–orbit coupled spin-1/2 Bose–Einstein condensate

We consider an effective two-dimensional Bose–Einstein condensate with some spin–orbit coupling (SOC) and a rotation term in an external harmonic potential. We find the striped state, and analyze the effects of SOC, the external potential, and the rotation frequency/direction on the profile and the stability of the striped state. Without the rotation term, the two spinor components exhibit striped pattern, and the numbers of stripes in the two components are always an odd–even or an even–odd. With the increase of the SOC strength, the number of stripes in both components increases, while the difference of the striped numbers is always one. After adding the rotation term, the profiles of the spinor components change qualitatively, and the change regulation of the striped numbers differs, while the difference of the striped numbers is still one. In addition, we find that the rotation direction only makes the striped state of the two spinor components exchange each other, though the clockwise and counterclockwise rotation directions are inequivalent with the presence of SOC. Such regulation is different from the previous study. And the rotation frequency gives rise to the transition from the striped state to a mixture of the striped state and vortex state. Furthermore, we prove the stability of these states by the evolution and linear stability analysis.     

Photonic nanoarchitectures in butterflies and beetles: valuable sources for bioinspiration

Nature began developing photonic nanoarchitectures millions of years before humankind. Often, in the living world, color is a communication channel that may influence the chance of the individual surviving as well as the chance to reproduce. Therefore, natural color‐generating structures are highly optimized by many millennia of evolution. In this review, a survey is presented of the development of natural photonic crystal‐type nanoarchitectures occurring in butterflies and beetles from the standpoint of physics and materials science, covering the past ten years. One‐, two‐, and three‐dimensional structures are reviewed, emphasizing the role that disorder, or irregularity, may play in natural nanoarchitectures to achieve certain visual effects. The characterization, modeling methods, and rapidly growing number of bioinspired or biomimetic applications are discussed.     

基于改进天牛群算法优化的BP神经网络的入侵检测

针对传统BP神经网络的入侵检测中,BP神经网络模型存在容易陷入局部最优、收敛速度慢、初始值随机性较大等缺点,本文提出改进天牛群算法(Beetle Swarm Optimization,BSO)用于优化BP神经网络的权值与阈值,并采用可变的感知因子及导向性的学习策略,以增强算法跳出局部最优的能力,提升算法全局寻优能力。利用天牛群算法群体智能的特点,提高BP神经网络的收敛速度。并将天牛群优化的BP神经网络模型应用于入侵检测,仿真实验结果表明优化后的BP神经网络模型能够显著提高模型的收敛速率和对入侵数据的检测率,降低误报率。     

基于改进BAS-BPNN的Stewart平台位姿正解

针对目前Stewart平台位姿正解难以兼顾高精度和实时性的问题，设计出满足平台精度要求且运算时间尽量少的正解方法是研究并联平台运动学的关键。首先建立平台运动学反解模型求得BP神经网络（BPNN）的训练集，然后调整步长策略和引入惯性权重改进天牛须搜索算法（improved BAS，IBAS），提升算法性能，通过改进后的算法优化BP神经网络的初始权阈值，反复学习训练至最小误差，最后选取一段位姿变化进行仿真验证。仿真结果表明，IBAS-BPNN正解法在选取的100组位姿点中的最大位姿误差小于0.6%，运算时间为31.9 s，通过与遗传算法(GA)、粒子群算法(PSO)和天牛须搜索算法(BAS)优化的BP神经网络模型进行对比分析，验证了IBAS-BPNN模型在精度和运算速度方面具有优越性。算法模型有效提升了Stewart平台位姿正解的精度和实时性，为平台的运动学和控制提供了理论和方法参考。