A hybridization of cuckoo search and particle swarm optimization for solving optimization problems

A new hybrid optimization algorithm, a hybridization of cuckoo search and particle swarm optimization (CSPSO), is proposed in this paper for the optimization of continuous functions and engineering design problems. This algorithm can be regarded as some modifications of the recently developed cuckoo search (CS). These modifications involve the construction of initial population, the dynamic adjustment of the parameter of the cuckoo search, and the incorporation of the particle swarm optimization (PSO). To cover search space with balance dispersion and neat comparability, the initial positions of cuckoo nests are constructed by using the principle of orthogonal Lation squares. To reduce the influence of fixed step size of the CS, the step size is dynamically adjusted according to the evolutionary generations. To increase the diversity of the solutions, PSO is incorporated into CS using a hybrid strategy. The proposed algorithm is tested on 20 standard benchmarking functions and 2 engineering optimization problems. The performance of the CSPSO is compared with that of several meta-heuristic algorithms based on the best solution, worst solution, average solution, standard deviation, and convergence rate. Results show that in most cases, the proposed hybrid optimization algorithm performs better than, or as well as CS, PSO, and some other exiting meta-heuristic algorithms. That means that the proposed hybrid optimization algorithm is competitive to other optimization algorithms.