PARALLEL CUCKOO SEARCH FOR GLOBAL OPTIMIZATION

The cuckoo search (CS) has become acceptable worldwide as one of the most efficient metaheuristic optimization search techniques applied to various real-world problems. Since the first appearance, many variants of the CS have been developed to improve its search performance. In this paper, the newest modified version of the CS named the parallel cuckoo search (PCS) is proposed for running on a single-CPU platform. Based on the original CS by using the random process drawn from the Levy distribution, the proposed PCS contains the CS as its search core. In the PCS algorithm, the partitioning strategy (PS) is conducted to divide an entire search space into many sub-search-spaces for each CS. The sequencing strategy (SS) is employed to organize the search units to run one-by-one on a single iteration/generation. Also, the discarding strategy (DS) is applied to discarding some unlikely to be successful CS. To perform its search performance, the proposed PCS is tested against ten selected benchmark optimization problems compared with the original CS. As experimental results, the proposed PCS performs more efficient in global optimization of ten selected benchmark optimization problems with higher success rates, less search generations and less search times than the original CS.