Hybridizing Cuckoo Search with Naked Mole-rat Algorithm: Adapting for CEC 2017 and CEC 2021 Test Suites

Cuckoo search (CS) has proved its worth and is one among the most competitive algorithm for numerical optimization. In order to improve its exploitation properties, this paper presents the hybridization of CS with a recently introduced naked mole-rat algorithm (NMRA). The major modification include
$i$
) new techniques based on barebones for global and NMRA for local search respectively are devised, ii) simulated annealing based mating factor for enhanced exploitation iii) an oscillating switch probability to balance between exploration and exploitation, and iv) shrinking population size reduction is used to minimize the computational burden. Apart from that, division of generations and population is also employed. The proposed mutation adaptive CS with MNRA (MaCN) is tested on CEC 2017 and CEC 2021 numerical benchmarks. From the experimental and statistical results, it can be said that MaCN is highly competitive with respect to MVMO, SaDN, JADE, SHADE, CV1.0, CSsin and CVnew algorithms.