【摘要】 神经网络算法可以建立属性参数与预测目标之间的高度非线性映射,遗传算法有选择适者生存的特点,神经网络算法与遗传算法构成的混合波阻抗反演具有精度更高,收敛速度更快的优势。本文简述了BP神经网络算法和遗传算法的基本原理,用理论模型比较了BP人工神经网络、遗传算法与混合波阻抗反演各自的特点,用实际地震资料检验了混合波阻抗反演的实用性。更多还原

【Abstract】 The BP neural network could develop the nonlinear mapping between the attribute parameters and the goal forecast.The genetic algorithm(GA) could preserve the best results.The impedance inverse method combined the BP neural network and GA has the advantages of high precision and fast converging.The basic principle of the BP neural network and GA is briefly introduced.The characteristics of the BP neural network,GA and the combined inverse are compared by using the theoretic model and the practicability of the combined impedance inverse is verified by the actual seismic data.更多还原