Details

Autor(en) / Beteiligte

• Farnad, Behnam
• Jafarian, Ahmad
• Baleanu, Dumitru

Titel

A new hybrid algorithm for continuous optimization problem

Ist Teil von

• Applied Mathematical Modelling, 2018-03, Vol.55, p.652-673

Ort / Verlag

New York: Elsevier Inc

Erscheinungsjahr

2018

Link zum Volltext

Quelle

EBSCOhost Business Source Ultimate

Beschreibungen/Notizen

• •An efficient hybrid method based on the GA, PSO and SOS algorithms.•The proposed method improves the natural selection process in GA by using the PSO and SOS.•The experiment based on the data clustering and benchmark function shows the better performance of the proposed method. This paper applies a new hybrid method by a combination of three population base algorithms such as Genetic Algorithm (GA), Particle Swarm Optimization (PSO) and Symbiotic Organisms Search (SOS). The proposed method has been inspired from natural selection process and it completes this process in GA by using the PSO and SOS. It tends to minimize the execution time and in addition to reduce the complexity. Symbiotic organisms search is a robust and powerful metaheuristic algorithm which has attracted increasing attention in recent decades. There are three alternative phases in the proposed algorithm: GA, which develops and selects best population for the next phases, PSO, which gets experiences for each appropriate solution and updates them as well and SOS, which benefits from previous phases and performs symbiotic interaction update phases in the real-world population. The proposed algorithm was tested on the set of best known unimodal and multimodal benchmark functions in various dimensions. It has further been evaluated in, the experiment on the clustering of benchmark datasets. The obtained results from basic and non-parametric statistical tests confirmed that this hybrid method dominates in terms of convergence, execution time, success rate. It optimizes the high dimensional and complex functions Rosenbrock and Griewank up to 10−330 accuracy in less than 3 s, out-performing other known algorithms. It had also applied clustering datasets with minimum intra-cluster distance and error rate.