Skip to Main content Skip to Navigation
Theses

Adaptation de la métaheuristique des colonies de fourmis pour l'optimisation difficile en variables continues. Application en génie biologique et médical.

Abstract : The ant colony algorithms are inspired by the collective behaviours observed in ant colonies and aim at solving difficult optimization problems.

The first approach to conceive some metaheuristics for the continuous optimization using this metaphor concist in creating a multi-agent system. We thus propose an "interacting ant colony" algorithms (CIAC). The second approach describe these metaheuristics as some methods manipulating a sampling of a probability density. We thus propose an "estimation of distribution" algorithm (CHEDA).

In line with the adaptive memory programming concept, our algorithms are hybridicized with a Nelder-Mead local search (HCIAC). We have afterward adapted this method to continuous dynamic problems (DHCIAC), and proposed a new coherent benchmark.

Our algorithms are finally applied as part of the automatization of the follow-up of eye injury.
Document type :
Theses
Complete list of metadatas

https://tel.archives-ouvertes.fr/tel-00093143
Contributor : Johann Dreo <>
Submitted on : Tuesday, September 12, 2006 - 5:45:28 PM
Last modification on : Wednesday, September 4, 2019 - 1:52:08 PM
Long-term archiving on: : Monday, April 5, 2010 - 11:39:50 PM

Identifiers

  • HAL Id : tel-00093143, version 1

Collections

Citation

Johann Dréo. Adaptation de la métaheuristique des colonies de fourmis pour l'optimisation difficile en variables continues. Application en génie biologique et médical.. Autre [cs.OH]. Université Paris XII Val de Marne, 2003. Français. ⟨tel-00093143⟩

Share

Metrics

Record views

469

Files downloads

18641