Title :
Imperialist competitive algorithm for natural crack shape reconstruction from ECT signals
Author :
Siquan Zhang ; Yu Liu ; Chang Yin ; Chuan Qi
Author_Institution :
Dept. of Electr. & Autom., Shanghai Maritime Univ., Shanghai, China
Abstract :
This paper presents a method to solve the problem of natural crack shape reconstruction from eddy current testing signals by means of imperialist competitive algorithm (ICA). ICA is a new meta-heuristic optimization and stochastic search strategy which is inspired from socio-political phenomenon of imperialistic competition. In order to evaluate the efficiency on solving crack shape inversion problem, ICA is compared with some heuristic algorithms such as genetic algorithm, particles swarm optimization and ant colony optimization. The reconstructed results verified the efficiency of neural network based forward model and the promising of imperialist competitive algorithm in crack shape inversion.
Keywords :
ant colony optimisation; crack detection; eddy current testing; genetic algorithms; image reconstruction; neural nets; particle swarm optimisation; shape recognition; stochastic processes; structural engineering computing; ECT signals; ant colony optimization; crack shape inversion problem; eddy current testing signals; genetic algorithm; heuristic algorithms; imperialist competitive algorithm; imperialistic competition; meta-heuristic optimization; natural crack shape reconstruction; neural network based forward model; particles swarm optimization; socio-political phenomenon; stochastic search strategy; Cost function; Eddy current testing; Genetic algorithms; Neural networks; Noise; Shape; Artificial neural network; Crack shape reconstruction; Eddy current testing; Imperialist Competitive Algorithm;
Conference_Titel :
Intelligent Control and Automation (WCICA), 2014 11th World Congress on
DOI :
10.1109/WCICA.2014.7053548
