Gestion dynamique des activités des chariots cavaliers sur un terminal portuaire à conteneurs en environnement incertain : approche par intelligence collective

Gaëtan Lesauvage (1)

1 LITIS - Laboratoire d'Informatique, de Traitement de l'Information et des Systèmes

The CALAS project consists in a laser measure system allowing to localize precisely straddle carriers in a container terminal. The information given by such a tool makes an optimization possible. A box terminal is an open system subject to dynamics, in which many events can occur. Among others, they concern container arrivals and departures. We aim to optimize the straddle carrier handling in order to improve the terminal management. The exact state of the system is unknown. The optimization process that we try to build must be fail-safe and adaptive. In this context, we propose an approach using a meta-heuristic based on Ant Colony to resolve the problem of assigning missions to straddle carriers. We built a simulator which is able to test and to compare different scheduling policies.

Mots clés: intelligence collective; colonies de fourmis colorées; graphe dynamique; optimisation multi-critères; problème dynamique de ramassage et de livraison avec fenêtre de temps

Domaines: Intelligence artificielle [cs.AI]

Gaëtan Lesauvage. Gestion dynamique des activités des chariots cavaliers sur un terminal portuaire à conteneurs en environnement incertain : approche par intelligence collective. 2009, http://majecstic2009.univ-avignon.fr/. ⟨hal-00485787⟩