Cuckoo search algorithm with deep search

• Published in: 2017 3rd IEEE International Conference on Computer and Communications (ICCC) pp. 2241 - 2246
• Main Authors: Zefan, Cai, Xiaodong, Yang
• Format: Conference Proceeding
• Language: English
• Published: IEEE 01.12.2017
• Subjects: bionic swarm optimization algorithm; Convergence; cuckoo search algorithm; deep search; Heuristic algorithms; Indexes; Mathematical model; opposition-based learning; paralleled algorithm; Search problems; Simulation
• DOI: 10.1109/CompComm.2017.8322934

Cuckoo search algorithm (CS) is a kind of bionic swarm optimization algorithm, which is simple and convenient. Although it has obvious advantages, it cannot converge to the optimal solution when dealing with high dimensional complex problems, so its global search ability needs to be improved. In this paper, opposition-based learning (OBL) strategy and local enhanced search are introduced to improve the basic CS. In the selection stage, an opposition-based swarm (ob-swarm) is generated on the basis of the original swarm. If an old cuckoo individual from the original swarm may be discarded, three cuckoo individuals are randomly selected from the ob-swarm to produce a new individual. The old one will be replaced by the new one if the new one is better. At the end of each generation, a potential optimal solution will be searched for locally around the current global optimal solution in the evolution direction. This local searching operation can make up the problem that the search step may be not appropriate. The simulation results show that the improved algorithm improves the global search ability, convergence speed and convergence precision of the algorithm.