Iteratively re‐weighted and refined least squares algorithm for robust inversion of geophysical data

Abstract

A robust metric of data misfit such as the ℓ1‐norm is required for geophysical parameter estimation when the data are contaminated by erratic noise. Recently, the iteratively re‐weighted and refined least‐squares algorithm was introduced for efficient solution of geophysical inverse problems in the presence of additive Gaussian noise in the data. We extend the algorithm in two practically important directions to make it applicable to data with non‐Gaussian noise and to make its regularisation parameter tuning more efficient and automatic. The regularisation parameter in iteratively reweighted and refined least‐squares algorithm varies with iteration, allowing the efficient solution of constrained problems. A technique is proposed based on the secant method for root finding to concentrate on finding a solution that satisfies the constraint, either fitting to a target misfit (if a bound on the noise is available) or having a target size (if a bound on the solution is available). This technique leads to an automatic update of the regularisation parameter at each and every iteration. We further propose a simple and efficient scheme that tunes the regularisation parameter without requiring target bounds. This is of great importance for the field data inversion where there is no information about the size of the noise and the solution. Numerical examples from non‐stationary seismic deconvolution and velocity‐stack inversion show that the proposed algorithm is efficient, stable, and robust and outperforms the conventional and state‐of‐the‐art methods.

Gravity Inversion Method to Produce Compact and Sharp Images Using L0-norm Constraint with Auto-adaptive Regularization and Combined Stopping Criteria

“…As pointed out in previous works (Farquharson & Oldenburg, 2004;Gholami & Aghamiry, 2017) instead of using a constant value of ℓ, dynamic re-adjustment throughout the iterative scheme might be a superior approach. Taking this into account, in the present work ℓ is updated in each iterative step.…”

Surface-Wave Extraction Based on Morphological Diversity of Seismic Events

“…To sparsely represent surface waves, Equation (3) is solved to achieve high-resolution LRT in the spectral bandwidth of surface waves by the iteratively reweighted least squares (IRLS) algorithm [27].…”

“…What causes this phenomenon, "mode kissing", is the non-negligible effect of the reflections within this range of frequencies and velocities. The picked dispersion curves based on the amplitude and the continuity of dispersive energy are shown in Figure 5c, where the second higher mode of frequencies of [25][26][27] Hz is misidentified as the third higher mode. However, the surface-wave dispersive energy on Z-component seismic data is not severely influenced by the reflections from the deep reflectors according to Hu et al [19].…”

An Optimized Newton-Raphson Algorithm for Approximating Internal Rate of Return

“…A more efficient way to estimate is by using a root-finding algorithm an iterative calculation scheme to approximate a single, isolated root of a function f(IRR), where the root IRR is a solution of the equation f(IRR) = 0 [12]. IRR is the most widely-used method in measuring the feasibility of a project or investment [4]- [6]. It is one of the tools that helps an intelligent enterprise in their decision making processes, which may help them in realizing certain goals [9].…”

“…However, [19] stated that, though its convergence is very fast and the number of the digits doubles in every iteration, which helps us in reaching our tolerance of errors quickly, the guarantee of convergence is still ambiguous and there are instances when division by zero occurs, and it is a considerable disadvantage of Newton-Raphson method, and suggested that the initial guess must be chosen very close to the true root [6], [7], [11]. It also requires an initial guess value from the user, which gives a high probability of non-convergence when the user's guess is far from the true root.…”