A novel meta-heuristic algorithm based on candidate cooperation and competition
Abstract Traditional meta-heuristic algorithms are often inspired by natural phenomena or biological behaviors, while relatively few are based on human social behavior. Moreover, existing algorithms inspired by human social behavior often suffer from premature convergence and getting trapped in loca...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Nature Portfolio
2025-07-01
|
| Series: | Scientific Reports |
| Subjects: | |
| Online Access: | https://doi.org/10.1038/s41598-025-08894-3 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Abstract Traditional meta-heuristic algorithms are often inspired by natural phenomena or biological behaviors, while relatively few are based on human social behavior. Moreover, existing algorithms inspired by human social behavior often suffer from premature convergence and getting trapped in local optima. To address these limitations, we propose a novel metaheuristic algorithm called the Candidates Cooperative Competitive Algorithm (CCCA), which is inspired by distinctive human social behaviors and designed for continuous optimization problems. CCCA consists of two main stages: self-study and mutual influence among candidates. The mutual influence stage includes various cooperative behaviors, such as one-on-one and many-to-one assistance, collaborative discussions among outstanding candidates, and targeted support for average candidates. Additionally, it incorporates competitive mechanisms, including contests among top-performing candidates and elimination strategies. We apply CCCA to solve 23 classical test functions, comparing them with PSO, FA, CSA, HMS, ICA, TLBO, and BSO. The results demonstrate that CCCA outperforms the compared algorithms, achieving optimal solutions in 9 functions. The convergence trends indicate that CCCA has a strong ability to escape local optima in 7 unimodal and multimodal functions. Statistical analysis using the Mann–Whitney U test confirms that CCCA achieves significant performance improvements in over 90% of the test cases. We also compare CCCA with recently developed human-inspired algorithms and obtain similarly competitive results. These findings further underscore the feasibility and robustness of CCCA. Moreover, its successful application to the capacity allocation problem highlights its practical effectiveness and superiority. |
|---|---|
| ISSN: | 2045-2322 |