원문정보
초록
영어
This paper presents an improved clustering algorithm with Ant Colony optimization (ACO) based on dynamical pheromones. Pheromone is an important factor for the performance of ACO algorithms. Two strategies based on adaptive pheromones which improved performance are introduced in this paper. One is to adjust the rate of pheromone evaporation dynamically, named as P , and the other is to adjust the strength of pheromone dynamically, named as Q . Two evaluation indices, Precision and Recall, are chosen to validity the improvement strategies. Numerical simulations demonstrate that the two strategies on pheromone can achieve better performance than basic ant colony algorithm and clustering algorithm with ant colony based on best solution kept.
목차
1. Introduction
2. Clustering Algorithm with ACO based on Dynamic Pheromone
2.1. Ant Colony System
2.2. Clustering algorithm with ACO
2.3. The strategy of dynamical pheromones
2.4. Clustering algorithm with ant colony optimization based dynamical pheromones
3. Simulation Experiment
3.1. Datasets and parameters setting
3.2. Comparison of results
3.3. Discussion
4. Conclusion and Future Work
Acknowledgements
References