Application of improved krill herd algorithms to inverse radiation problems

A nature-inspired metaheuristic krill herd (KH) algorithm was introduced to inverse transient radiative problems for the first time. Three improved KH (IKH) algorithms were developed to accelerate convergence in the later stage of the searching process. Radiative source terms were then estimated in a parallel slab of participating medium to investigate the performance of the proposed algorithms. Results showed the superiority of the three IKH algorithms to the original one. Additionally, the extinction coefficient and scattering albedo in a parallel slab with short pulse laser incident were retrieved using the IKH algorithms. Radiative properties cannot be estimated accurately with measurement errors when the media were optically thick and scattering dominated in the inverse transient radiative problems. Hence, adopting a shorter incident pulse was suggested to solve this problem. Consequent numerical simulations indicated that radiative properties can be retrieved accurately even with measurement errors. Finally, the proposed IKH algorithms were applied to a 2D media to obtain the optical properties of the inclusion. (C) 2015 Elsevier Masson SAS. All rights reserved.