Magnetic anomaly inversion through the novel barnacles mating optimization algorithm

Area of Science:

• Geophysics
• Inverse Problems
• Optimization Algorithms

Background:

• Geophysical inverse problems are often ill-posed and non-unique, requiring regularization and prior information when using local optimizers.
• Local search algorithms are sensitive to initial model selection, while global optimization methods offer broader exploration without this constraint.
• Global optimization and metaheuristic algorithms are increasingly adopted in geophysics for parameter estimation.

Purpose of the Study:

• To introduce and apply the Barnacles Mating Optimizer (BMO), a novel global optimizer, to the geophysical inverse problem of magnetic anomaly interpretation.
• To evaluate BMO's performance against a standard algorithm (sPSO) in terms of convergence, stability, robustness, and accuracy.
• To assess the effectiveness of a second moving average (SMA) scheme for regional anomaly removal and to perform post-inversion uncertainty analysis.

Main Methods:

• Application of the Barnacles Mating Optimizer (BMO), a bio-inspired global optimization algorithm.
• Testing on simulated magnetic anomalies and real-world data from chromite, uranium deposits, and a Mesozoic dike.
• Utilizing a second moving average (SMA) scheme for regional anomaly removal and conducting uncertainty assessment analyses.

Main Results:

• BMO demonstrated an ability to scan the model parameter space more extensively than standard Particle Swarm Optimization (sPSO).
• The algorithm consistently approached the unique global minimum without compromising accuracy or stability.
• The SMA scheme was certified for regional anomaly elimination, and uncertainty analyses confirmed the reliability of BMO's solutions.

Conclusions:

• The Barnacles Mating Optimizer (BMO) is a competitive and effective tool for geophysical inverse problems, particularly magnetic anomaly interpretation.
• BMO offers advantages in exploring the solution space and achieving reliable results compared to traditional methods like sPSO.
• The study recommends the use of BMO for model parameter estimation in various geophysical applications.