Generalized Nets as Tools for Modeling of the Ant Colony Optimization Algorithms

Ant Colony Optimization (ACO) has been used successfully to solve hard combinatorial optimization problems. This metaheuristic method is inspired by the foraging behavior of ant colonies, which manage to establish the shortest routes to feeding sources and back. We discuss some possibilities for describing of the ACO algorithms by Generalized Nets (GNs), that help us deeply to understand the processes and to improve them. Very important is that the GNs are expandable. For example, some of the GN places can be replaced with a new GN. Thus we can include procedures to improve the search process and the achieved results or various kind of estimations of algorithm behavior.