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.