Format

Send to

Choose Destination
Neural Comput. 2017 Apr;29(4):1103-1123. doi: 10.1162/NECO_a_00946. Epub 2017 Feb 9.

Parameter Estimation of Nonlinear Systems by Dynamic Cuckoo Search.

Author information

1
College of Meteorology and Oceanography, PLA University of Science and Technology, Nanjing, 211101, China liaoqixiang2013@126.com.
2
College of Meteorology and Oceanography, PLA University of Science and Technology, Nanjing, 211101, China sq20112@163.com.
3
College of Meteorology and Oceanography, PLA University of Science and Technology, Nanjing, 211101, China mask1000@126.com.
4
College of Meteorology and Oceanography, PLA University of Science and Technology, Nanjing, 211101, China shq20111@163.com.

Abstract

In order to address with the problem of the traditional or improved cuckoo search (CS) algorithm, we propose a dynamic adaptive cuckoo search with crossover operator (DACS-CO) algorithm. Normally, the parameters of the CS algorithm are kept constant or adapted by empirical equation that may result in decreasing the efficiency of the algorithm. In order to solve the problem, a feedback control scheme of algorithm parameters is adopted in cuckoo search; Rechenberg's 1/5 criterion, combined with a learning strategy, is used to evaluate the evolution process. In addition, there are no information exchanges between individuals for cuckoo search algorithm. To promote the search progress and overcome premature convergence, the multiple-point random crossover operator is merged into the CS algorithm to exchange information between individuals and improve the diversification and intensification of the population. The performance of the proposed hybrid algorithm is investigated through different nonlinear systems, with the numerical results demonstrating that the method can estimate parameters accurately and efficiently. Finally, we compare the results with the standard CS algorithm, orthogonal learning cuckoo search algorithm (OLCS), an adaptive and simulated annealing operation with the cuckoo search algorithm (ACS-SA), a genetic algorithm (GA), a particle swarm optimization algorithm (PSO), and a genetic simulated annealing algorithm (GA-SA). Our simulation results demonstrate the effectiveness and superior performance of the proposed algorithm.

PMID:
28181879
DOI:
10.1162/NECO_a_00946

Supplemental Content

Full text links

Icon for Atypon
Loading ...
Support Center