Dimension-by-dimension enhanced cuckoo search algorithm for global optimization

Abstract

Cuckoo search (CS) algorithm is an efficient meta-heuristic algorithm that has been successfully applied in many fields. However, the algorithm uses the whole updating and evaluating strategy on solutions. For solving multi-dimensional optimization problems, solutions with partial dimension evolution may be discarded due to mutual interference among dimensions. Therefore, this strategy may deteriorate the quality solution and convergence rate of algorithm. To overcome this defect and enhance the algorithm performance, a dimension-by-dimension enhanced CS algorithm is proposed. In the global explorative random walk, the improved algorithm uses the dimension-by-dimension updating and evaluating strategy on solutions. This strategy combines the updated values of each dimension with the values of other dimensions into a new solution. In addition, a greedy strategy is adopted to accept new solution and the search center is set as the current optimal solution. The proposed algorithm was tested on fourteen well-known benchmark functions. The numerical results show that the improved algorithm can effectively enhance the quality solution and convergence rate for the global optimization problems.