An efficient social-driven educational competition optimizer for numerical optimization

C Qian and XY Cai, CLUSTER COMPUTING-THE JOURNAL OF NETWORKS SOFTWARE TOOLS AND APPLICATIONS, 28, 1024 (2025).

DOI: 10.1007/s10586-025-05728-9

The Educational Competition Optimizer (ECO) is a human-based meta- heuristic algorithm. Like other meta-heuristic algorithms, ECO has drawbacks such as an imbalance between exploration and exploitation, an uneven population distribution, and a propensity to enter local optimum states. This work suggests the Social Driven Educational Competition Optimizer (SDECO), which combines two crucial strategies, to get beyond these restrictions. First, this work proposes a new search phase: the university phase. This phase improves population diversity and avoids falling into local optima. Second, a social driven learning strategy is integrated into the ECO through an adaptive framework, which enhances the search capability of the ECO through the co-leading of dominant groups and dominant individuals. Together, these strategies improve the convergence speed, stability and optimization ability, ensuring an effective balance between exploration and exploitation. The CEC2017 test suite is used to evaluate the suggested algorithm's strategy efficacy and parameter sensitivity. The CEC2020 and CEC2022 test suites are used to compare SDECO with the basic and enhanced algorithms, respectively. The results show the superior performance of SDECO in terms of convergence speed, stability and global optimization, achieving an average ranking of 1.50 (CEC2020) and 1.75 (CEC2022). Compared to basic ECO, SDECO achieves 94.8%(CEC2017), 90.0%(CEC2020) and 95.8%(CEC2022) performance improvement in the three test suites, respectively. In addition, we further validate the ability of SDECO to solve real-world optimization problems. According to the results, SDECO achieved an average ranking of 1.250. To sum up, SDECO is a promising meta-heuristic algorithm variant.

Return to Publications page