Quantum-behaved PSO with neighborhood strategies for solving NDET inverse problems

The paper studies the efficiency of a QPSO (Quantum-behaved Particle Swarm Optimization) algorithm enhanced with neighborhood strategies to solve a NDET (Non-Destructive Electromagnetic Testing) inverse problem, formulated as an optimization problem. Two different neighborhood strategies are analyzed and compared with the classic QPSO, one based on disjoint subswarms, and the other based on informants. Among the neighborhood strategies, some problem specific local search techniques are applied to speed up the inversion process. The NDT inverse problem consists in the reconstruction of 3D partially conductive cracks with uniform conductivity, conductivity smaller than the conductivity of the tested specimen.