LAUSR

In this paper, a clustering cuckoo search optimization (CCSO) is proposed. Different from the randomly generated step size in CSO, the step size in CCSO is generated by a clustering mechanism, and the value is updated according to the average fitness value difference between each cluster and the whole swarm, thereby improving the searching balance between exploration and exploitation of each solution. The effectiveness of CCSO has been validated by six typical benchmark functions and economic load dispatch problems with 6, 10, 13, 15 and 40 generators. The results of CSO and CCSO are displayed and compared in aspects of convergence rate, objective function value and robustness. Moreover, the influences of parameters as step size $$\delta $$ δ , solution number P , egg abandon fraction $$p_a$$ p a and cluster number K are all analyzed comprehensively in this study. The conclusion is that, in all the tested cases, CCSO behaves much more competitive than CSO under the same parameter setting conditions.