Abstract
Swarm-based algorithms are a powerful family of optimization techniques inspired by forming flocks, colonies and swarms. In this paper, the swarm intelligence concepts of particle swarm optimization (PSO), which is an effective and reliable algorithm, gravitational search algorighm (GSA) and cuckoo search algorithm (CKA), which are recently developed meta-heuristic algorithms, were analyzed. The numerical optimization problem solving successes of these algorithms were compared by testing about 50 different benchmark functions. Numerical results revealed that CKA exhibited the highest performance in solving various nonlinear functions, while PSO and GSA produced better results on multimodal and multivariable problems. The obtained results also showed that GSA and CKA supplied more robust than the PSO. The CKA is essentially expressed by Lévy flight and allows more efficient in exploring the search space as its step length is much longer in the long run, leading to better performace in convergence, precision and robustness.
