Staged hybrid genetic search for seismic data imaging

Seismic data interpretation problems are typically solved using computationally intensive local search methods which often result in inferior solutions. Here, a traditional hybrid genetic algorithm is compared with different staged hybrid genetic algorithms on the geophysical imaging static corrections problem. The traditional hybrid genetic algorithm used here applied local search to every offspring produced by genetic search. The staged hybrid genetic algorithms were designed to temporally separate the local and genetic search components into distinct phases so as to minimize interference between the two search methods. The results show that some staged hybrid genetic algorithms produce higher quality solutions while using significantly less computational time for this problem.<<ETX>>