Neural Network with Adaptive Immune Genetic Algorithm for Eddy Current Nondestructive Testing

For eddy current nondestructive testing (ECNDT), an immune genetic algorithm (IGA) is presented, which can overcome the disadvantages of genetic algorithm (GA), such as possibility of being trapped on locally minimum value and prematurity convergence. Moreover, crossover and mutation operators are selected by adaptive algorithm to overcome prematurity. Compared with GA, the convergence precision and generalization of IGA are improved remarkably