摘要
一种基于弱T-范数和弱S-范数的神经元,可以实现与、或和混合-并模糊逻辑运算,并且拥有较强的鲁棒性。将它所组成的神经网络运用到模糊推理系统中,不仅可以简化网络,实现模糊推理最基本的一致性要求,还可以控制在模糊推理过程中当规则发生摄动时对推理结果的影响程度。
The neurons based on a weak T-norm and the weak S-norm can be achieved with and, or and mixed-and fuzzy logic operations, and are robust. Based on the new neurons, a neural network is applied to fuzzy inference systems, which can not only simplify the network, satisfy the consistency principle of fuzzy inference, but also control the extent of the impact of the results of reasoning with the perturbation of rules in fuzzy inference.
出处
《计算机工程与科学》
CSCD
北大核心
2010年第3期148-150,共3页
Computer Engineering & Science
基金
湖南省教育厅科研基金资助项目(07A056)
关键词
神经元
神经网络
鲁棒性
摄动
控制
neurons
neural network
robustness
perturbation of rules
control
