混合智能优化地震反演

传统非线性反演方法存在着收敛效率低，有时会陷入局部极值的问题。为此，研究开发了融合粒子群优化算法和郭涛算法的混合智能优化地震非线性反演技术。粒子群优化算法对解的更新更具有目的性，收敛速度快；郭涛算法构造了多父体交叉重组算子并采用群体爬山策略，求解精度高；混合智能优化算法以粒子群优化算法为主体框架，融入郭涛算法的寻优机制。函数优化测试、理论模型试算和实际资料反演处理结果表明，混合智能优化算法具有求解效率高、全局寻优能力强的优点，适合解决复杂的地震反演问题。

Seismic inversion of hybrid intelligent optimization

The conventional non-linear seismic inversion methods are low in convergence and sometimes will result in the problem of local extremum. To solve these problems, we developed a non-linear seismic inversion of hybrid intelligent vptimiza-tion with the particle warm optimization algorithm and the Guo's algorithm integrated. The particle warm algorithm is charac-terized by higher objective in solution upgrading and rapid convergence, while the Guo's algorithm constructs a multi-parent combination crossover and adopts a colony mountain climbing search strategy, thus is high in accuracy of solution. The hybrid intelligent algorithm uses the particle warm optimal algorithm as the framework and takes advantage of the optimization mecha-nism of the Guo's algorithm. We performed function optimization test, trial calculation with the theoretical model and real data inversion processing. The results show that this method has the advantages of high efficiency and strong capacity of global opti-mization, thus is suitable for complex seismic inversion.