This algorithm optimizes the comparison between experimental data and numerically calculated, or simulated data, from randomly generated conductivity profiles. The algorithm looks for the best fit between measured and simulated data, and from this fit, conductivity profile is estimated. In metallic multilayer flat pieces, good approximation of profile have been found for thickness in the order of the standard skin depth.

Conductivity in each descretization cell, is related to the average conductivity within the cell. Thickness and conductivity are usually determined calibrating adequately the standard eddy current instruments. Although, recovering of multilayer conducting profiles, requires the solution of the related inverse reported and deterministic numerical algorithms as well as stochastic ones, have been applied to practical problems. In the present work, a simple of inverting electromagnetic data is applied to axial symmetry situation in order to recover thin multilayered profiles for metallic materials. Data recording requires a single radial scanning and digitizing the output (amplitude and phase) from absolute air coils or Hall effect detectors. With the implemented procedure, conductivity and thickness of metallic sheets, up to 3 mm, have been estimated in the range of 3 to 50 kHz.

Non magnetic materials like copper, aluminum, plumb and tin were used in this experiments, including non conducting interlayers. The average conductivity was successfully employed as the parameter required for determining the jump, according to the visiting distribution in the so called fast simulated annealing (FSA) or Cauchy machine algorithm.