节点文献
基于变尺度寻优和遗传搜索技术的模糊神经网络全局学习算法
A New Global Learning Algorithm for Fuzzy Neural Networks Based on Modified Quasi-Newton Method and Genetic Searching Techniques
【摘要】 本文给出了一种将改进的拟牛顿算法与具有新型交配方式和可变变异概率的遗传算法相结合的全局寻优算法,用以搜索模糊神经网络误差函数的全局最小点。在拟牛顿算法中采用了一种基于模糊推理和组合插值技术的线性搜索算法,可以用较少步数得到二次型置信区间内的全局最优步长,确保局部寻优具有快速收敛特性。关于算法分析的定理证明了这种混合算法对于紧致集内的权向量构成的任意连续函数能依概率1收敛于全局极小值。计算机仿真结果表明,本算法既具有较快的收敛速度,又具备良好的全局收敛特性。
【Abstract】 In this paper,a new global optimizing algorithm that combines the modified Quasi-Newton method and the improved genetic algorithm is proposed to find the global minimum of the total error function of a fuzzy neural network.A global linear search algorithm based on fuzzy logic and combinative interpolation techniques is developed in the modified Quasi-Newton rnodel.It is shown that this algorithm ensures convergence to a global minimum with probability 1in a compact region of a weight vector space.The results of computer simulations also reveal that this algorithm has a better convergence property and the times of global search decrease obviously.
【Key words】 Quasi-Newton method; Combinative interpolation; Trust region; Improved genetic algorithm; Fuzzy neural network; Global optimizing algorithm; Convergence with probability;
- 【文献出处】 电子学报 ,ACTA ELECTRONICA SINICA , 编辑部邮箱 ,1996年11期
- 【分类号】O242.23
- 【被引频次】9
- 【下载频次】148