【作者】 李医民；

【导师】 胡寿松；

【作者基本信息】 南京航空航天大学 ， 控制理论与控制工程， 2004， 博士

【摘要】 本文以复杂非线性生态系统为研究背景,利用模糊数学建立了生物系统中生态位及相关概念的数学模型。研究了生物系统中的混沌、同步、分岔等非线性动力学行为,初步构建了生态系统的数学理论框架。同时,将生态系统中存在自适应、鲁棒性和稳定性等生物特性抽象出来,通过数学方法形成生物特性融入到控制系统和优化方法中,提出了一类具有生物特性的模糊T-S控制方法和模糊容错控制方法。建立了基于生物特性的优化方法。提出了扩展的Type-2模糊系统,并建立了具有冗余结构和冗余补充特性的生物系统的模糊数学模型。最后,将基于生物特性的控制与优化方法应用到生态系统中,研究生物系统的非线性行为、预测与控制。首先,以生态位(Niche)和生物群落为基本单元,抽象研究了生态位理论、生态系统的边缘效应、建立了生态位与群落的模糊数学关系原理。提出了相似优先选择竞争原理、物种间生态位的选择、竞争及演化原理,证明了高斯竞争定理。另外,本文所建立的生态位模型是目前常用的三种生态位模型的统一形式,这为生态位的实际应用带来方便。利用动力学方法,对生态系统进行非线性分析,研究了生态系统的混沌现象、物种竞争系统的同步行为以及生物系统的分形。提出了生物进化的混沌度量方法。其次,利用生态系统本身具备的自适应、自学习和自组织等特性,将生态系统的这种特性融入到遗传算法中,提出了基于生态位技术的遗传算法的动态改进方法,弥补了常规的静态改进方法,提高遗传算法的收敛性和稳定性。并将基于生态位技术的遗传算法应用于模糊系统的参数辨识和模糊规则的自动生成。最后,本文将生物系统生物进化发展的容错性能和鲁棒性与常规的控制方法相结合,将生态位作为模糊T-S模型的后件,建立了基于生物特性的模糊T-S模型,提出了基于生物冗余特性的模糊故障诊断方法。将这些方法应用混沌系统的控制、生物系统的控制和智能温室的控制及复杂系统的容错控制。更多还原

【Abstract】 Based on complex non-linear ecosystems, this paper use fuzzy mathematics to found the mathematical models of niche and its relative concepts of biological systems and study the non-linear dynamic behaviors, such as chaos, synchronization, and bifurcation of biological systems. Primarily, it forms the framework of mathematic theoretic studies. At the same time, by abstracting biological traits, such as salf-adaption, robustness and stadility, from the bio-systems, and melting biotechnology into control systems and optimization methods by mathematical methods, it proposes a type of fuzzy T-S control methods and fuzzy fault-tolerant control methods which has biological traits, and forming optimization methods based on biotechnology. At the end, this paper applies the control and optimization methods based on biological traits in studying the non-linear behaviors, prediction bio-cybernetics of ecosystems.At first, regarding niches and biological communities as basic elements, this paper studies the niche theory, marginal effects of ecosystems, the principle of fuzzy mathematical relation between niche and community, proposing similar prior choice competitive principle, the choice of niches between species and competitive and evolving principle, and verifying Guass competition theorem. Furthermore, since the niche model founded in this paper is the unification of the tree commonly used niche models, this makes it convenient to use the model practically.Using dynamic methods, it does nonlinear analysis to ecosystems, studying the chaotic phenomena of ecosystems, synchronous behaviors of species-competition systems and the fractals of bio-system, and giving a measure method of the bio-evolution chaotic degree. Second, this paper melts the traits the ecosystems have itself, such as self-adaptation, self-learn and self-organization, into genetic operation, and proposes a genetic operational dynamical improving method based on niche-technology, a supplement of the conventional motionless improving method, increasing the convergence ang stability of the genetic operation. And it uses the genetic operation based on niche-technology in the parameter identification of fuzzy systems and the self-generating of fuzzy regularity.At last, combining the fault-tolerant character and robustness in the biological evolution process of bio-systems with formal control methods, regarding niche as a rear parameter of fuzzy T-S system, this paper forms a fuzzy T-S system based on biological traits and proposes a fuzzy fault diagnosis method based on biological redundancy. It applies all these methods it gets in the control of chaotic systems, ecosystems and the intellect greenhouse and the fault-tolerant control of complex systems.更多还原