基于免疫进化规划的一种柔性神经模糊推理系统

A Flexible Neuro-Fuzzy Inference System Based on Immune Evolu tion ary Programming

该文基于泛逻辑学提出一种新颖的柔性神经模糊推理系统,用命题间的广义相关性和广义自相关性去解释系统推理模式的连续可变,以及命题真值的测量误差,以期实现真正的智能控制系统,并采用了将进化规划同生物免疫思想中的浓度机制及多样性保持策略相结合的免疫进化规划学习算法,自适应地学习系统参数。最后通过倒立摆的仿真实验体现了该推理系统的可用性和有效性。

Based on the principle of Universal Logics,this paper proposes a novel flexible neuro-fuzzy inference sys-tem,which uses general correlation and general self-correlation between propositions to explain the reasoning form to change in succession and the measuring errors of the membership grades of fuzzy proposition,respectively.And this sys-tem uses an immune evolutionary programming to learn adaptively the parameters of this system.Simulation results show that our approach is practical and effective.