Comparison of the performance of global optimization methods with perturbative methods in the estimation of ocean bottom properties from multifrequency data

Abstract

During 1988 a series of acoustic propagation experiments was conducted in the Hudson Canyon area off the coast of New Jersey. It included synthetic aperture experiments in which a source transmitting multiple tones was towed toward/away from a vertical array of receivers. By this means, a two-dimensional structure of the acoustic field was obtained at eight different frequencies. This data set is used to estimate the bottom acoustic properties using global optimization approaches such as simulated annealing and genetic algorithm and to compare their performance with that of perturbative methods, which because of their simplicity have some advantages over the global methods.