【作者】 陈建良；

【导师】 朱伟兴；

【作者基本信息】 江苏大学 ， 控制理论与控制工程， 2006， 硕士

【摘要】 传统控制是建立在系统的精确数学模型基础上的，而实际系统常常存在复杂性、非线性、时变性、不确定性等问题，难以获得精确数学模型。作为智能控制的一种新技术，模糊控制以其实现简单、鲁棒性强、不依赖精确数学模型等优点得到广泛应用。常规的模糊控制器设计由设计者通过总结专家的经验构成模糊控制规则。然而，在专家经验不完善，甚至个别经验不正确的情况下，得到的模糊规则不能完全反映整个控制系统的本质特征。在这些情况下，通过优化算法来优化模糊控制器的模糊规则，改善模糊控制器的性能，已经成为当前模糊控制应用的重要课题。 蚁群算法是近年来涌现出的一种新颖的优化算法，已被广泛且有效地应用到求解复杂的组合优化问题中。本文对蚁群算法进行了较为全面的分析研究，并针对基本蚁群算法的收敛慢、容易停滞等不足，对蚁群算法的选择策略和信息素更新机制进行了改进。采用确定性选择和随机选择相结合的方法改进了选择策略，增加了较优路径的被选概率，有效缩短了搜索时间；采用信息素浓度的最大值和最小值限制的方法改进了信息素更新机制，扩大了搜索空间。基于TSP的研究结果显示了本文所提出的改进算法的收敛速度快和搜索空间大的良好性能。模糊规则的确定是一个组合优化问题，本文尝试使用蚁群算法来优化得到模糊规则，并提出了基于蚁群算法的模糊控制器优化设计方法，实现模糊控制规则的自寻优。系统阶跃响应的研究表明了经自寻优得到的模糊控制器具有较好的动态性能和稳态性能。对寻优所得的模糊控制器的鲁棒性分析表明，该模糊控制器具有较强的鲁棒性，充分体现了模糊控制的优点。更多还原

【Abstract】 Traditional control is constituted based on the accurate mathematics model. But the real system has its properties such as complexity, non-linear, real time, indeterminate and so on. It is hard to obtain the accurate mathematics model. As a new technology of intelligent control, fuzzy control is widely used for its virtues such as easy realization, robust. The routine method of fuzzy controller design is to sum experts’ experience up to fuzzy control rules. However, the gained fuzzy rules can not reflect the essence characters of the whole control system when the experts’ experience is incomplete, and especially the individual experience is incorrect. In this circumstance, it is an important task of application in design of fuzzy controller to improve the capability of fuzzy controller by optimizing the fuzzy rules.Ant colony algorithm is a novel optimization algorithm welled up in recent years, and has been widely used to solve the complex combined optimization problem. the paper makes a full research to it, and offer an improved algorithm instead of the basic ants algorithm’s disadvantage such as convergence slowly and stagnancy easily, etc. The simulation of the research shows that this improved algorithm has a faster convergence speed and larger search space, and it also has a better result of parameter improvement. To confirm fuzzy rules is a issue of combined optimization problem. In this paper, ant colony algorithm is applied into the optimization design of fuzzy controller. A lot of simulation results indicate the feasibility and validity of the algorithm, and the optimized fuzzy controller has more dynamic, steadier and more robust capability.更多还原