Subspace approach to inversion by genetic algorithms involving multiple frequencies

Based on waveguide physics, a subspace inversion approach is proposed. It i s observed that the ability to estimate a given parameter depends on its se nsitivity to the acoustic wavefield, and this sensitivity depends on freque ncy. At low frequencies it is mainly the bottom parameters that are most se nsitive and at high frequencies the geometric parameters are the most sensi tive. Thus, the parameter vector to be determined is split into two subspac es, and only part of the data that is most influenced by the parameters in each subspace is used. The data sets from the Geoacoustic Inversion Worksho p (June 1997) are inverted to demonstrate the approach. In each subspace Ge netic Algorithms are used for the optimization - it provides flexibility to search over a wide range of parameters and also helps in selecting data se ts to be used in the inversion. During optimization, the responses from man y environmental parameter sets are computed in order to estimate the a post eriori probabilities of the model parameters. Thus the uniqueness and uncer tainty of the model parameters are assessed. Using data from several freque ncies to estimate a smaller subspace of parameters iteratively provides sta bility and greater accuracy in the estimated parameters.