Generalized non-linear elastic inversion with constraints in model and data spaces

摘要: SUMMARY Seismic data are non-linearly related to model parameters such as seismic velocities. However, inversion is usually considered in a linear approximation. Such techniques the Born were recently applied data. Non-linear more complicated and involves extensive calculations. Non-linear was developed frame work of an unconstrained optimization procedure. It uses priori information initial probability distribution functions spaces (This called ‘soft’ bounds). In this paper, we propose new technique for solving constrained non-linear inversion. This will allow us use not only terms bounds, but ‘hard’ bounds well (usually giving stable accurate solutions). Non-linear iterative procedure which dual transform at each iteration. A allows considering problem Lagrangian multipliers. The number multipliers equal available thus, significantly reduces dimension (this true underdetermined problems only). most important property that it consider problem. Another proper constraints incorporate small-wave numbers generalized shown conventional (unconstrained inversion) special case paper if truncation operator represented by identity matrix.