Abstract : Dynamic island models are population-based algo-rithms for solving optimization problems, where the individualsof the population are distributed on islands. The purpose of thisdynamic/adaptive management of the migrations is to send theindividuals to the most promising islands, with regards to theircurrent states. In this paper, we propose to investigate Q-Learningas migration policy.
https://hal.univ-angers.fr/hal-02709479 Contributeur : Okina Univ AngersConnectez-vous pour contacter le contributeur Soumis le : mercredi 13 octobre 2021 - 11:15:54 Dernière modification le : mercredi 3 novembre 2021 - 06:08:43 Archivage à long terme le : : vendredi 14 janvier 2022 - 18:46:43
Frédéric Lardeux, Frédéric Saubion, Jorge Maturana. Learning and Migration Processes for Island Models. 28th IEEE International Conference on Tools with Artificial Intelligence (ICTAI), 2016, San Jose, United States. ⟨hal-02709479⟩