【作者】 唐志航；

【导师】 俞立；

【作者基本信息】 浙江工业大学 ， 控制理论与控制工程， 2003， 硕士

【摘要】 模糊控制自提出至今已在理论和实际应用方面取得了很大的进展。这种方法以模糊规则的形式充分利用系统的局部信息和专家操作经验，通过模糊推理得到全局模型，从而成为解决复杂对象的建模和控制问题的一种有效方法。但是，在模糊控制器的设计过程中仍然存在两个瓶颈问题：一是模糊控制规则的选取和优化。模糊控制规则的选取主要依赖于该领域专家的先验知识和工人的操作经验，有较大的主观性，当系统很复杂或输入变量较多时，模糊控制规则的可能选取空间急剧加大，造成规则“爆炸”，而当控制系统采用典型的两输入一输出的线性控制规则时，控制规则的设计缺乏形象的指导；二是模糊子集和模糊隶属函数的确定。在模糊控制规则确定的情况下，模糊控制系统的性能由模糊变量的各个模糊子集的隶属函数来决定，这是多参数寻优问题，很难获得全局最优。 基于上述考虑，本文结合粗糙集理论、集对分析理论、滑动模态理论和遗传算法，研究了模糊控制器优化设计问题。论文主要内容如下： 1)分析了模糊控制规则抽取中存在的困难，用粗糙集理论和集对分析理论解决模糊控制中模糊控制规则的抽取和过滤问题，提出了一种新的规则提取方法，并给出了一个应用实例。结果表明，该方法对一大类复杂系统的模糊控制规则的优化具有一定的应用价值。 2)利用滑动模态的理论，用定量和定性的方法对典型的二阶模糊控制器的解析结构进行分析，利用模糊控制理论和滑动模态变结构理论的相似性，特别是滑动模态变趋近律控制的思想对模糊控制规则表进行形象的分析，并推导出模糊控制规则表解析结构表达式，利用此方法可进行模糊控制规则的优化设计，仿真实验结果验证了该方法的有效性。 3)将遗传算法与模糊控制相结合，提出了先使用模糊逻辑的思想进行交叉概率和变异概率的整定，再利用模糊遗传算法对模糊子集的划分进行寻优，可获得一个基于一定性能指标的次优或最优模糊控制器。以二阶系统为例进行了计算机仿真，研究结果表明这种方法是有效的。 最后，对全文工作进行了总结，分析了存在的问题和缺陷，并对今后的发展方向作了展望。更多还原

【Abstract】 It makes great development about the theory ,and application of Fuzzy Logic Control (FLC) in recent years. Because this approach can efficiently utilize system local information and expert operation experiences to get the general model by means of fuzzy reasoning, it resolves what’s meaning constructing model and controlling for complex plants. But, there still exists two bottlenecks in design of FLC: One is the choice and optimization of fuzzy rules. Because the choice of rules depends on expert operation experiences, it is subjective. While the control system is complex or inputs are much more, the probable selective space quickly becomes large so that it will cause "rules burst". Even while the control system uses linear fuzzy rules with two inputs and one output, the design of fuzzy rules is short of graphic guidance. The other is the choice of the fuzzy set partition and membership function. After gaining the fuzzy rules, the performance of FLC depends on membership function of each fuzzy set partition. On account of optimization of many parameters, acquiring general optimization is very difficult.Based on above discuss facts, this dissertation makes a study on the design of optimization of FLC using a combination of fuzzy logic, rough sets, sets pair analysis, sliding mode and genetic algorithms. The main contents are as follows:1) Both rough and fuzzy sets theories can be used to deal with imprecise and incomplete information. Both of them are often used to observe, test and reason about data. This chapter applies degree of connection in sets pair analyzed to rough sets theory. The relationship between rough-degree of connection and cardinal of rough lower/upper approximation is demonstrated. The difficulties of drawing fuzzy control rules are analyzed in this paper. A new method for drawing and filtering fuzzy control rules by using rough sets and sets pair analyzed is proposed. The result shows this method is quite effective in fuzzy rules optimization for a large class of complex systems.2) For a typical fuzzy controller with two inputs and one output, a method of analyzing the rule-base with the theory of sliding mode is presented when it useslinear fuzzy control rules. Both graphic and numerical analysis illustrate or demonstrate that typical FLC possesses properties of its linear control and sliding mode control. These dual features can help us understand the robustness of FLC systems. Simulation results show the effectiveness of the proposed method.3) To combine fuzzy logic control with genetic algorithm, On the one hand, fuzzy control can express the systemic information with non-linear and fuzzy knowledge; on the other hand, genetic algorithms can improve the systemic ability of self-study. In this chapter, fuzzy genetic algorithm is used to optimize the fuzzy set partitions and the fuzzy logic control method is used to adjust the probabilities of crossover and mutation. It can be obtained to the design of optimal or sub-optimal fuzzy logic controller. For an industrial process described by a second-order model, a computer simulation is conducted. The result shows the proposed method is effective.Finally, on the basis of the summarization of the research results in this dissertation, the future developments about FLC are discussed.更多还原