Evolutionary computation-based metaheuristic algorithms have been successfully applied to hard optimization problems. In this very active research area, one of the newest algorithms is a Cuckoo search metaheuristic capable of solving general Ndimensional, linear and nonlinear optimization problems. It is easy to understand and apply, and requires simple mathematical preprocessing. Ever since its foundation in 2009, Cuckoo search algorithm has drawn the attention of many researchers all over the world, resulting in a lot of variants of the basic algorithm with improved performance. This paper reviews the basic concepts of Cuckoo search algorithm, a survey of its major variants applied to solve diverse optimization problems in engineering sciences and the theoretical studies conducted on it so far. Also, it provides an overview of the significant engineering applications that have benefited from its powerful nature.

For many optimization problems of interest, the complexity of the problem and the computational constraints make it very difficult to find an exact solution within a reasonable time. In these cases, it is common practice to rely on the use of metaheuristic optimization algorithms that can find a good, maybe even the optimal solution in a reasonably short time (Yang, 2010). However, it does not guarantee about the quality of the found solution. The power and beauty of almost all recent metaheuristics optimization algorithms come from the capability of emulating the best features in nature, specifically biological systems evolved from natural selection over millions of years via two important characteristics, which are selection of the fittest and adaptation to the environment (Yang and Deb, 2009).

Blum and Roli (2003) gave two fundamental characteristics of metaheuristics, which were intensification and diversification. Intensification, also termed as exploitation, typically searches around the instant best solutions locally and intensively, while diversification, also termed as exploration, explores the search space globally and efficiently, often by large-scale randomization. The fine balance between these two is very significant to the overall efficiency and performance of an algorithm. There is a tradeoff whether the problem requires more intensification and less diversification or vice versa.

Electrical and Electronics Engineering Journal, Cuckoo search, Metaheuristic, Optimization.