基于改进萤火虫算法的波阻抗反演方法

Wave impedance inversion method based on improved firefly algorithm

为了解决在全局寻优搜索过程中,传统萤火虫算法(FA)求解精度低、后期收敛速度慢和易陷入局部极值等问题,笔者提出一种基于混沌搜索和动态步长相结合的萤火虫算法(CAFA)。在搜索初期,CAFA算法使用具有随机性、有界性和遍历性的混沌序列设置萤火虫种群初始位置;利用一种新型双曲线递减动态步长代替传统固定步长。经过上述改进提高了种群分布的均匀性、求解精度、收敛速度以及避免迭代后期出现反复振荡现象。将CAFA算法应用于地震资料波阻抗反演,利用理论模型论证算法可行性、有效性和抗噪性。实验结果表明CAFA算法具有较好反演效果。

In order to solve the low accuracy,slow convergence in late stages and easiness to fall into local optima and other issues of the traditional firefly algorithm(FA)in global optimization search process,the authors propose a firefly algorithm based on combined chaos search and dynamic step(CAFA).In the initial stage of the search,the CAFA algorithm uses a random,bounded and ergodic chaotic sequence to set the initial position of the firefly population;a new type of hyperbolic decreasing dynamic step size is used to replace the traditional fixed step size.These improvements enhance the uniformity of population distribution,solution accuracy,and convergence speed,and avoid the repeated oscillation phenomenon occurred at later stages.The theoretical models demonstrate the feasibility,effectiveness and noise immunity of the CAFA algorithm,when it is applied to seismic wave impedance inversion.Experimental results show that the CAFA inversion algorithm has a good inversion effect.