2009 | OriginalPaper | Buchkapitel
Revisiting the Auto-Regressive Functions of the Cross-Entropy Ant System
verfasst von : Laurent Paquereau, Bjarne E. Helvik
Erschienen in: Self-Organizing Systems
Verlag: Springer Berlin Heidelberg
Aktivieren Sie unsere intelligente Suche, um passende Fachinhalte oder Patente zu finden.
Wählen Sie Textabschnitte aus um mit Künstlicher Intelligenz passenden Patente zu finden. powered by
Markieren Sie Textabschnitte, um KI-gestützt weitere passende Inhalte zu finden. powered by
The Cross-Entropy Ant System (CEAS) is an Ant Colony Optimization (ACO) system for distributed and online path management in telecommunication networks. Previous works on CEAS have enhanced the system by introducing new features. This paper takes a step back and revisits the auto-regressive functions at the core of the system. These functions are approximations of complicated transcendental functions stemming from the Cross-Entropy (CE) method for stochastic optimization, computationally intensive and therefore not suited for online and distributed operation. Using linear instead of hyperbolic approximations, new expressions are derived and shown to improve the adaptivity and robustness of the system, in particular on the occurrence of radical changes in the cost of the paths sampled by ants.