【摘要】 针对模糊控制中量化因子难以确定以及无先验知识情况下模糊控制器结构和参数难以得到问题,提出了一种无量化因子和不依赖专家经验知识的模糊直接逆控制方案,克服了传统模糊规则基的得到过分依赖专家知识的弱点。控制器的规则基不采用规则的三角形形式,而是采用与实际更加接近的无规则三角形,规则基的得到不依赖已知经验,而是根据系统输入输出信息,利用进化规划离线对控制器规则基的结构和参数同时优化。实际控制中利用优化的规则基对系统进行控制,实时控制算法的运算时间仅为一组规则基的计算时间,大大加快了系统的响应速度,仿真结果验证了算法的有效性。更多还原

【Abstract】 In order to solve problems of gaining factors, structure and parameters in fuzzy control processing, with study of lots of method of fuzzy control, a new method of fuzzy direct inverse control without factors and prior experience is proposed. With combination of conventional method and new thoughts, the given method can solve two kinds of difficulties: one kind lies in determining numerical factors in conventional fuzzy control, and the other lies in determining structure and parameters of controller in practice processing. The new method overcomes the demerit of conventional methods in which rule base is largely derived by premised conditions and fuzzy rule base of controller used in conventional fuzzy control is systematic, in the paper fuzzy rule base is unsystematic triangle, it close to practice, and moreover, deriving fuzzy rule bases is based on input and output information of given system. Evolutionary programming is used in offline training to optimize structure and parameter of controller抯 rule base simultaneously. In practical control, we use the optimal base to control the given system. The time of practical control algorithm is only equal to time of one rule base, so the response speed becomes faster than conventional methods. The effectiveness is proved by simulation with nonlinear system.更多还原