A Bio-Inspired Swarm Approach to 5000-Dimensional Optimization(papers.ssrn.com)
papers.ssrn.com
A Bio-Inspired Swarm Approach to 5000-Dimensional Optimization
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5593065
https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5593065
DHCS is a bio-inspired metaheuristic designed for high-dimensional and complex optimization problems, addressing limitations of conventional approaches like PSO or Genetic Algorithms.
Key features:
Dynamic clustering & adaptive roles: Each agent autonomously decides its behavior while maintaining swarm coherence.
Periodic synchronization: Ensures global coordination without sacrificing exploration.
Scalability: Tested on a 5000-dimensional Ackley function with superior convergence and robustness.
Efficiency: Reduces computational overhead while outperforming standard methods.
Versatility: Applicable to engineering design, supply chain optimization, ML hyperparameter tuning, and financial modeling.
This paper not only formalizes the DHCS framework but also presents a comprehensive experimental evaluation demonstrating its effectiveness in high-dimensional and dynamic environments.
I’d love feedback from the community, especially from those working in metaheuristics, swarm intelligence, and large-scale optimization problems.