细菌觅食优化算法的研究进展

细菌觅食优化算法是近年来发展起来的,基于大肠杆菌觅食行为模型的一种新型智能算法。它具有对初值和参数选择不敏感、鲁棒性强、简单易于实现,以及并行处理和全局搜索等优点。但其在应用过程中存在精度不够高、收敛速度不够快的缺点。文中首先对细菌觅食优化算法的基本原理及操作流程进行介绍,并概述了国内外学者在这一领域的研究现状,接着分析了算法三大主要操作存在的问题,然后探讨了算法的改进和应用,最后分析了算法未来的研究方向。     

Ecological consequences of chemically mediated prey perception

To locate food, mobile consumers in aquatic habitats perceive and move towards sources of attractive chemicals. There has been much progress in understanding how consumers use chemicals to identify and locate prey despite the elusive identity of odor signals and the complex effects of turbulence on chemical dispersion. This review highlights how integrative studies on behavior, fluid physics, and chemical isolation can be fundamental in elucidating mechanisms that regulate species composition and distribution. We suggest three areas where further research may yield important ecological insights. First, although basic aspects of stimulatory molecules are known, our understanding of how consumers identify prey from a distance remains poor, and the lack of studies examining the influence of distance perception on food preference may result in inaccurate estimation of foraging behavior in the field. Second, the ability of many animals to find prey is greatest in unidirectional, low turbulence flow environments, although recent evidence indicates a trade-off in movement speed versus tracking ability in turbulent conditions. This suggests that predator foraging mode may affect competitive interactions among consumers, and that turbulence provides a hydrodynamic refuge in space or time, leading to particular associations between predator success, prey distributions, and flow. Third, studies have been biased towards examining predator tracking. Current data suggest a variety of mechanisms prey may use to disguise their presence and avoid predation; these mechanisms also may produce associations between prey and flow environments. These examples of how chemical attraction may mediate interactions between consumers and their resources suggest that the ecology of chemically mediated prey perception may be as fundamental to the organization of aquatic communities as the ecology of chemical deterrence.     

Postbite elevation in tongue-flicking rate by an iguanian lizard,Dipsosaurus dorsalis

The herbivorous iguanid lizardDipsosaurus dorsalis exhibited PETF (postbite elevation in tongue-flicking rate), an increase in tongue-flicking rate after experimental removal from the mouth of food that had been bitten. This was demonstrated by a significantly higher tongue-flick rate after having bitten food than in three experimental conditions controlling for responses to the experimental setting, sight of food, and mechanical disturbance caused by the experimental removal of food from a lizard's mouth. As in most other families of lizards, PETF was brief, occurring only during the first minute. Lizards are divided into two major suprafamilial taxa, Iguania and Scleroglossa, consisting of carnivorous species characterized by two major foraging modes, ambush and active, and of herbivores and omnivores. PETF is absent in the two families of carnivorous iguanian lizards studied that are ambush foragers but present in three families of scleroglossan lizards that are active foragers. However, PETF is absent in the two species studied in a scleroglossan family, Gekkonidae, which forages by ambush, and present in an iguanian herbivore, as reported herein. We propose that the presence or absence of PETF, in addition to its phylogenetic determinants, is adaptively adjusted to foraging mode.     

基于细菌迁徙的自适应果蝇优化算法

针对果蝇优化算法的早熟收敛问题,提出了一种新的基于细菌迁徙的自适应果蝇优化算法。该算法在运行过程中根据进化停滞步数的大小自适应地引入细菌迁徙操作,提高算法跳出局部极值的能力;并且对每个个体根据适应值大小赋予不同的自适应迁徙概率,避免了迁徙可能带来的解退化的问题。对几种经典函数的测试结果表明,新算法具有更好的全局搜索能力,在收敛速度、收敛可靠性及收敛精度上比果蝇优化算法有较大的提高。     

改进细菌觅食算法求解车间作业调度问题*

针对细菌觅食算法（BFOA）求解高维优化问题时容易陷入局部最优和早熟的问题,引入自适应步长及差分进化算子,并将改进算法用于车间作业调度问题（JSP）中。求解时,设计了一种编码转换方案,从而无须修改BFOA运算规则即可实现对JSP的寻优;同时,采用空闲时间片段优化策略降低了调度问题的复杂性。仿真实验表明,该算法能够跳出局部最优,避免了早熟的问题,调度结果优于原始细菌觅食算法和离散粒子群算法。     

蜂群算法研究综述*

蜂群算法是一种模仿蜜蜂繁殖、采蜜等行为的新兴的群智能优化技术,近几年备受研究者关注。初步探讨了蜂群算法的理论基础,详细论述了基于蜜蜂繁殖行为和采蜜行为的两类蜂群算法的生物学机理及其最常见算法的应用研究情况,并分析比较了遗传算法、蚁群算法、粒子群算法和蜂群算法的优缺点、适用范围及性能。最后,总结了现有蜂群算法存在的问题,并指出其未来的研究方向。     

Predatory behavior inTupinambis teguixin (Sauria: Teiidae). I. Tongue-flicking responses to chemical food stimuli

Black tegu lizards (Tupinambis teguixin) have the ability to detect food odors and discriminate between them and nonfood odors. This was tested by offering chemical stimuli on cotton-tipped applicators to the animals. Stimuli were from two plant and two animal species known to be principal items in these lizards' diets, demineralized water as an odorless control, and eaude-cologne as an odorous control lacking feeding or social importance. Tongueflick attack score, latency to attack, preattack tongue-flicks, and number of attacks were analyzed. The results clearly demonstrated that this species responds to chemical food stimuli, but does not respond to odorless nonfood stimuli. Responses differed among food types. There were no sex differences. These results are in agreement with the prediction that lizards having forked tongues and an active foraging mode rely on chemical cues for feeding.     

广义菌群优化算法

为提高菌群优化算法的性能,将群体聚集机制和自适应策略集成到趋药性操作中,取消聚集操作,构造出新的趋化操作,在趋化循环中引入自适应扩散机制,提高其克服“早熟”的能力,重新定义健康度,减少计算复杂性,得到了一种新的群体智能优化方法—广义菌群优化算法(GBFO, Generalized Bacterial Foraging Optimization)。通过10个复杂Benchmark函数的计算进行算法性能测试,并与几个典型的算法进行了实验比较,结果表明,GBFO算法在搜索能力和稳定性、求解质量和效率等方面优于其他典型算法的比率分别达到80%~90%,70%~80%,验证了该算法的优越性能。     

用于指数熵多阈值分割的改进细菌觅食算法

针对图像多阈值分割中阈值搜索是有序正整数规划的特点,提出了一种用于指数熵多阈值分割的改进细菌觅食优化(Improved Bacterial Foraging Optimization,IBFO)算法。首先,将标准的细菌觅食优化(Standard Bacterial Foraging Optimization,SBFO)算法的趋化算子改成动态趋化算子以增强趋化操作的自适应性;然后,将SBFO中的迁徙算子替换成混合随机和动态的迁徙算子,将迁徙过程划分为两个阶段,第一阶段为随机迁徙,目的是增强全局搜索能力,第二阶段为动态局部迁徙,目的是提高局部搜索能力;随后,丢弃SBFO中的感应机制以便加快运行速度;最后,将IBFO算法进一步修改以满足有序正整数规划的要求,并将其应用于指数熵多阈值分割方法中。图像分割实验结果表明,与SBFO,MBFO和IPSO算法相比,提出的IBFO方法不仅优化效果更好,而且运行速度更快。     

Social cues influence foraging in dwarf hamsters (Phodopus campbelli).

In 2 experiments, dwarf hamsters (Phodopus campbelli) were trained to find palatable foods in an open field. The location of each food patch remained the same throughout each experiment, and only 1 food was available per day. Once subjects had been trained to find each food in its unique location, they progressed to a testing phase in which subjects’ mates were allowed to eat and hoard the food that was available in the open field each day. The foods that subjects’ mates brought back to the home cages then served as discriminative stimuli signaling which food could be obtained in the open field. Subjects generally approached the patch containing the food hoarded by their mates, suggesting that dwarf hamster burrows could function as information centers. (PsycINFO Database Record (c) 2010 APA, all rights reserved)