基于遗传算法的参数优化在控制器设计中的应用

本文详述了遗传算法作为一种随机搜索算法在控制器设计参数优化中的应用。从遗传算法基本原理入手，结合工程实际，论述了遗传算法在PID控制器设计、鲁棒控制器设计、最优控制、系统参数辨识、模糊逻辑控制系统和神经网络控制中的应用成果。讨论了影响遗传算法的因素，并提出了改进的策略。     

航空板结构强度的边界元法

边界元即边界积分方程方法,可方便地用来解板弯曲问题。本文在局部极坐标下建立公式,采用具有较高协调性的函数插值方案,利用外点技术完全避免边界奇异积分。列举的一系列算例表明该方法输入数据少、机时省、精度高,是板结构强度分析的一个有效方法。     

Microbial life in a liquid asphalt desert

Pitch Lake in Trinidad and Tobago is a natural asphalt reservoir nourished by pitch seepage, a form of petroleum that consists of mostly asphaltines, from the surrounding oil-rich region. During upward seepage, pitch mixes with mud and gases under high pressure, and the lighter portion evaporates or is volatilized, which produces a liquid asphalt residue characterized by low water activity, recalcitrant carbon substrates, and noxious chemical compounds. An active microbial community of archaea and bacteria, many of them novel strains (particularly from the new Tar ARC groups), totaling a biomass of up to 10(7) cells per gram, was found to inhabit the liquid hydrocarbon matrix of Pitch Lake. Geochemical and molecular taxonomic approaches revealed diverse, novel, and deeply branching microbial lineages with the potential to mediate anaerobic hydrocarbon degradation processes in different parts of the asphalt column. In addition, we found markers for archaeal methane metabolism and specific gene sequences affiliated with facultative and obligate anaerobic sulfur- and nitrite-oxidizing bacteria. The microbial diversity at Pitch Lake was found to be unique when compared to microbial communities analyzed at other hydrocarbon-rich environments, which included Rancho Le Brea, a natural asphalt environment in California, USA, and an oil well and a mud volcano in Trinidad and Tobago, among other sites. These results open a window into the microbial ecology and biogeochemistry of recalcitrant hydrocarbon matrices and establish the site as a terrestrial analogue for modeling the biotic potential of hydrocarbon lakes such as those found on Saturn's largest moon Titan.