A Study of Hybridized Smell Agent Symbiotic Organism Search in Congress on Evolutionary Computation Functions

This paper presents a study of the Smell Agent Symbiotic Organism Search (SASOS) hybrid algorithm. SASOS is developed from bioinspired Smell Agent-Based Optimization(SAO) and Symbiosis Organism Search (SOS) algorithms. Bioinspired algorithms often lack a balance between speed and accuracy to achieve optimal performance efficiency and a global search for the best solution. To address these challenges, the algorithm reduces the imbalance between diversification and intensification in bioinspired algorithms to improve the search for global optima. SASOS performance was evaluated in sixteen selected Congress on Evolutionary Computation (CEC) functions using Aggregative Best Counts (ABC) compared to the regular SAO and SOS algorithms. For an advanced performance comparison, the convergence study was carried out on each CEC function to assess the fitness of the algorithms based on the Desirable Convergence Goal (DCG). Evaluation results using 50 iterations have shown that SASOS performed better withABCof56.25%than the SAO and SOS algorithms with ABC of 28.12% and 15.63%, respectively, in standard benchmark functions. Furthermore, in the convergence study, 1000 iterations were superimposed for each algorithm on the CEC functions. The convergence results showed that SASOS obtained the best DCG of 58.83%compared to SOS and SAO with DCG of 25.00% and 16.67%, respectively. These results made the performance of the hybrid SASOS uniquely different from other similar approaches.This is because the hybrid SASOS satisfactorily balanced the diversification and intensification phases in the bioinspired SAO and SOS algorithms. The eligible characteristics of the hybrid SASOS with respect to ABC and DCG showed its compatibilityand significance forvarious engineering optimizationapplications