Asynchronous distributed solution of large scale nonlinear inversion problems Original Research Article

In this paper we consider large scale nonlinear least squares problems whose objective functions are very expensive to evaluate, and whose Jacobian matrices are generally ill-conditioned and have an almost block diagonal structure. We prove the convergence of an asynchronous iteration to solve this kind of problems under standard assumptions. The method is naturally parallelizable and thus is applicable to practical inversion problems, as those arising in seismic exploration for hydrocarbons.