Ant Colony Optimization
A Nonlinear Multi-agent System designed for Swarm Intelligence: the Logistic MAS
SASO '07 Proceedings of the First International Conference on Self-Adaptive and Self-Organizing Systems
Hi-index | 0.00 |
Ant algorithms are usually derived from a stochastic modeling based on some specific probability laws. We consider in this paper a full deterministic model of "logistic ants" which uses chaotic maps to govern the behavior of the artificial ants. We illustrate and test this approach on a TSP instance, and compare the results with the original Ant System algorithm. This change of paradigm--deterministic versus stochastic--implies a novel view of the internal mechanisms involved during the searching and optimizing process of ants.