Exploration in stochastic algorithms: An application on MAX-MIN Ant System

In this paper a definition of the exploration performed by stochastic algorithms is proposed. It is based on the observation through cluster analysis of the solutions generated during a run. The probabilities associated by an algorithm to solution components are considered. Moreover, a consequent method for quantifying the exploration is provided. Such a measurement is applied to MAX-MIN Ant System. The results of the experimental analysis allow to observe the impact of the parameters of the algorithm on the exploration.