Simulated Annealing Algorithm Coupled With a Deterministic Method for Parameter Extraction of Energetic Hysteresis Model

The existing methods for the parameter identification of energetic hysteresis model have limitations of slow convergence and low accuracy. Aiming at this problem, a robust and efficient hybrid algorithm that combines simulated annealing (SA) method with the Levenberg–Marquardt (L–M) technique is proposed. Since SA has the ability to avoid traps in local minima, it is used to get to the zone nearby the global optimal point in the initial search period. Then based on the commutation criterion, optimizing process is transferred to the second search period using the normalized L–M algorithm, in which a normalization of the model sensitivity function is conducted to improve the convergence. The normalized L–M algorithm takes the current best solution of SA as its initial parameters, and converges rapidly toward the global minimum. The simulation and experimental results show that the proposed hybrid algorithm can lead to a considerable reduction in computation resources and provide accurate solution.