Next: Iteratively Reweighted Least Squares Up: Ji: Conjugate Guided Gradient Previous: Ji: Conjugate Guided Gradient

# CG method for LS and IRLS Inversion

Most inversion problems start by formulating the forward problem, which describes the forward operator, , that transforms the model vector to the data vector :

 (1)
In general, the measured data may be inexact, and the forward operator may be ill-conditioned. In that case, instead of solving the above equation directly, different approaches are used to find an optimum solution for a given data .The most popular method is finding a solution that minimizes the misfit between the data and the modeled data .The misfit, or the residual vector, , is described as follows:

 (2)
In least-squares inversion, the solution is the one that minimizes the squares of the residual vector as follows:

 (3)

Iterative solvers for the LS problem search the solution space for a better solution in each iteration step, along the gradient direction (in steepest-descent algorithms), or on the plane made by the current gradient vector and the previous descent-step vector (in conjugate-gradient algorithms). Following Claerbout 1992, a conjugate-gradient algorithm for the LS solution can be summarized as follows:



iterate {

} ,

where the subroutine cgstep() remembers the previous iteration descent vector, , where i is the iteration step, and determines the step size by minimizing the quadrature function composed from (the conjugate gradient) and (the previous iteration descent vector), as follows Claerbout (1992):

Notice that the gradient vector () in the CG method for LS solution is the gradient of the squared residual and is determined by taking the derivative of the squared residual (i.e. the L2-norm of the residual, , with respect to the model ):

 (4)