Directional Cuckoo Search Algorithm

Cuckoo search algorithm is a meta-heuristic population-based stochastic optimization algorithm inspired by the behavior of brood parasitism and lévy flight habit. Due to the simple concepts, fast convergent speed, and easy implementation, it has been widely applied into many areas. In this thesis, we discuss the global search and local search patterns for cuckoo search algorithm.In the standard version of cuckoo search algorithm, the ratio between local search and global search capability is associated with cuckoo’s performance significantly. In other words, the local search capability plays an important role for those cuckoos with worse performances, as well as cuckoos with better performances prefer to the global search. To improve the performance, a new local search pattern is designed to increase the influence of the best position obtained by the entire population, and this variant is called oriented cuckoo search algorithm. To test the performance, CEC2013 benchmark test suits are employed and compared with other five algorithms, simulation results show the validity.For oriented cuckoo search algorithm, the global search capability is mainly dominated by the hybrid distribution combined with Lévy distribution and Gaussian distribution. Therefore, in this thesis, different distributions are employed to test the influence, such as Cauchy distribution, Gaussian distribution, Exponential distribution, Uniform distribution and Lévy distribution, total ten possible hybrid distributions are discussed and tested on CEC2013 benchmark test suits, simulation results show the hybrid distribution between Lévy distribution and Cauchy distribution achieves the best performance. Furthermore, oriented cuckoo search algorithm with this hybrid distribution is also compared with four recent algorithms, simulation results show it achieves the best performance with Friedman test and Wilcoxon test.Finally, oriented cuckoo search algorithm is applied to improve the performance of DV-Hop localization algorithm. In this application, a new method to estimate unknown node coordinates is designed. Experiments show DV-Hop algorithm with cuckoo search algorithm significantly improves the positioning accuracy of unknown nodes and stability of the algorithm.