An Improved Firefly Algorithm Enhanced by Negatively Correlated Search Mechanism
Shi Wang, Xiao Yang, Zonghui Cai, Lin Zou, Shangce Gao · 2018
Firefly algorithm (FA) is inspired by natural phenomena and it is an effective optimizer for solving complex problems. However alike other swarm intelligent algorithms, FA also suffers from the premature convergence problem. To further improve the search effectiveness and alleviate this issue, the hybridization of different algorithms has shown to be a promising research direction. In this paper, we for the first time propose a hybrid algorithm, called NCFA by combing the firefly algorithm with the negatively correlated search. The characteristics of firefly algorithm make population diversity decline rapidly, which is more likely to lead to premature convergence. The core of the negatively correlated (NC) search is considered to be a special diversity control strategy. Experimental results based on CEC2017 benchmark functions demonstrate the superiority of such hybridization, and the diversity analysis of population also verify its rationality.