Population Diversity-Aided Adaptive Cuckoo Search

Cuckoo search (CS) is an evolutionary algorithm based on Levy flight distribution (LFD). The exploration and exploitation of CS are largely dependent on a probabilistic parameter \(P_a\) , the value of which ranges between 0 and 1. A lesser \(P_a\) value promises higher exploration, while a higher value leads to higher exploitation. The original CS algorithm has a \(P_a\) value fixed to 0.25. A controlled varying \(P_a\) suggests a richer exploration-exploitation balance. In this article, we have presented a version of the algorithm with a dynamic \(P_a\) value guided by real-time population diversity and have cited the results pertaining to improved efficiency and accuracy when tested upon fifteen CEC-15 benchmark functions.