| _version_ | 1866901213916692480 |
|---|---|
| author | Zhang, Jincheng |
| author_facet | Zhang, Jincheng |
| contents | <p><span>Swarm optimization algorithms are widely studied due to their lack of gradient information requirement and good adaptability to complex non-convex problems. However, most existing methods still rely on immediate fitness-driven and single global optimum guidance mechanisms, which are prone to premature convergence, rapid loss of search diversity, and excessive preference for unstable solutions in high-dimensional, multimodal, noisy, or dynamic environments. To address these limitations, this paper proposes a novel swarm optimization method—the cognitive yellow-throated marten optimization algorithm. This method no longer treats the optimization process as a simple random search or heuristic iteration, but models it as a dynamic decision-making system driven by counterfactual utility evaluation, cognitive differentiation subgroup cooperation, and game consistency constraints. The algorithm introduces a counterfactual utility function to characterize the difference in potential gains between individual actions and not taking that action, achieves self-organized division of labor in exploration, evaluation, and utilization within the group through a cognitive differentiation mechanism, and extends the optimization objective from a single-point optimum to a set of solutions with robustness and evolutionary stability by constructing a stable and survivable solution set. Theoretical analysis shows that this method has stronger stability and continuous search capability in perturbed environments, providing a new research paradigm for swarm optimization</span>.</p> |
| format | Recurso digital |
| id | zenodo_https___doi_org_10_5281_zenodo_18357887 |
| institution | Zenodo |
| language | |
| publishDate | 2026 |
| publisher | Zenodo |
| record_format | zenodo |
| spellingShingle | Yellow-throated Marten Optimization Algorithm Zhang, Jincheng <p><span>Swarm optimization algorithms are widely studied due to their lack of gradient information requirement and good adaptability to complex non-convex problems. However, most existing methods still rely on immediate fitness-driven and single global optimum guidance mechanisms, which are prone to premature convergence, rapid loss of search diversity, and excessive preference for unstable solutions in high-dimensional, multimodal, noisy, or dynamic environments. To address these limitations, this paper proposes a novel swarm optimization method—the cognitive yellow-throated marten optimization algorithm. This method no longer treats the optimization process as a simple random search or heuristic iteration, but models it as a dynamic decision-making system driven by counterfactual utility evaluation, cognitive differentiation subgroup cooperation, and game consistency constraints. The algorithm introduces a counterfactual utility function to characterize the difference in potential gains between individual actions and not taking that action, achieves self-organized division of labor in exploration, evaluation, and utilization within the group through a cognitive differentiation mechanism, and extends the optimization objective from a single-point optimum to a set of solutions with robustness and evolutionary stability by constructing a stable and survivable solution set. Theoretical analysis shows that this method has stronger stability and continuous search capability in perturbed environments, providing a new research paradigm for swarm optimization</span>.</p> |
| title | Yellow-throated Marten Optimization Algorithm |
| url | https://doi.org/10.5281/zenodo.18357887 |