Hybrid Genetic Autostatics: New Approach For Large-Amplitude Statics With Noisy Data

摘要: Large amplitude statics or noisy data cause probms in residual methods that use a master trace ck correlation peaks. A solution may not be found eoduces coherent reflectors the stacked section jlution suffer from cycle skips. combination of vo new complimentary approaches, genetic algorithms Id waveform steepest ascent, into hybrid method can vercome these pitfalls and produce more optimal soluons. The danger traces undergoing skips atic solutions correspond to local maxima stack awer objective function. problem avoiding linirna requires global search mechanism Ich as algorithm. However, mechanisms excessive computation cost rhen they start with collection random models. qaveform based ascent is an efficient stable pproach for finding near tarting model, This approach used here I provide initial high-graded population enetic algorithm speed ccasional iterations volution. Waveform “climbs” unction rather than picking Repeated pplication hill climbing allows one find leak without pre-determined bias. In addition, we he cross-correlation every CDP Ither avoid algorithmic problems related trace. large set correlations jrovides vastly over-determined system equations which helps stabilize inversion, particularly lseful data. To efficiently determine slopes massive crosscorrelations, accelerated iterative matrix inver. sion borrowed tomographic inversion. serves model fol testing inverse could useful othel applications seismology.