论文部分内容阅读
提出一种新颖的基于ＤＮＡ的进化算法（ＤＮＡ－ＥＡ）来自动设计一类Ｔａｋａｇｉ Ｓｕｇｅｎｏ（ＴＳ）模糊控制器．ＴＳ模糊控制器采用带有线性规则后项的ＴＳ模糊规则，连续输 入模糊集，Ｚａｄｅｈ模糊逻辑和常用的重心反模糊器．ＴＳ模糊控制器被证明是带有可变增 益的非线性ＰＩ控制器．ＤＮＡ－ＥＡ被用于自动获取ＴＳ模糊规则，并同时优化模糊规则前 项和后项中的设计参数．ＤＮＡ－ＥＡ采用由生物ＤＮＡ结构启发得到的ＤＮＡ编码方法来编 码模糊控制器的设计参数．在ＤＮＡ－ＥＡ中，引入了受微生物进化现象启发的基因转移和细 菌变异操作．另外，也引入了基于 ＤＮＡ遗传操作的框构变异操作． ＤＮＡ编码方法非常适 合于复杂知识的表达，基于基因水平的遗传操作也很容易引入到ＤＮＡ－ＥＡ中．染色体的长 度是可变的，且可插入或删除部分碱基序列。作为示例，给出了采用ＤＮＡ－ＥＡ来自动设计 ＴＳ模糊控制器用于控制一类非线性系统的方法．ＤＮＡ－ＥＡ能自动地构造模糊控制器．计 算机仿真结果表明，ＤＮＡ－ＥＡ是有效的，且优化得到的模糊控制器是满意的．
A novel DNA-based evolutionary algorithm (DNA-EA) is proposed to automatically design a Takagi Sugeno (TS) fuzzy controller. TS fuzzy controller adopts the TS fuzzy rule with the linear rule of the latter term, continuously inputs the fuzzy set, Zadeh fuzzy logic and the commonly used center of gravity anti-fuzzy device. TS fuzzy controller turns out to be a nonlinear PI controller with variable gain. DNA-EA is used to automatically obtain TS fuzzy rules and simultaneously optimize the design parameters in the pre- and post-fuzzy rules. DNA-EA uses DNA coding methods inspired by biological DNA structures to encode the design parameters of the fuzzy controller. In DNA-EA, gene transfer and bacterial mutation motifs inspired by microbial evolution are introduced. In addition, a framework mutation operation based on DNA genetic manipulation has also been introduced. DNA coding method is very suitable for the expression of complex knowledge, gene-based genetic manipulation is also very easy to introduce into DNA-EA. The chromosome length is variable, and some of the base sequences can be inserted or deleted. As an example, a method of using DNA-EA to automatically design a TS fuzzy controller for controlling a class of nonlinear systems is given. DNA-EA automatically constructs a fuzzy controller. Computer simulation results show that DNA-EA is effective, and the fuzzy controller optimized is satisfactory.