【摘要】 提出一种利用非单点模糊系统建模的短波通信频率参数预测模型。该模型以非单点模糊系统为平台,利用梯度下降算法和遗传算法构成的混合并行学习算法训练建立推理规则。通过对电离层实测数据进行仿真试验,并与神经网络方法、单点模糊系统方法进行比较,结果证明该模型具有预测精度高、收敛速度快、全局收敛性好、抗噪声能力强等优点。更多还原

【Abstract】 This paper presents a prediction model of frequency parameters of short-wave communication based on non-singleton fuzzy system.The system parameters of the model were adjusted by a new parallel learning algorithm.The new learning algorithm was composed of gradient descent algorithm and genetic algorithm.The model is simulated and compared with other nonlinear methods. Some standout excellences are shows,such as higher precision of prediction,faster speed of convergence,better global convergence and better anti-jamming capability.更多还原