Implémentation d'une méta-heuristique embarquée pour résoudre le problème d'ordonnancement dans un atelier flexible de production

Abstract : Scheduling problems in the industrial sector are the most studied optimization problems. However, resolution approaches are not suited to a dynamic and increasingly uncertain environment. The work presented in this thesis is aiming for solving the Flexible Job Shop Scheduling Problem (FJSSP) without and with uncertainties. We propose a Particle Swarm Optimization (PSO) algorithm as well as two multi-agent architectures based on PSO called MAPSO1 and MAPSO2 to solve the FJSSP problem without uncertainty. Testing our algorithms on different benchmarks has shown the effectiveness of the obtained results. We are also proposing to use embedded systems for the deployment of an intelligent and cooperative optimization system based on PSO. Thus the MAPSO2 architecture has been distributed on a physically distributed system comprising two embedded systems. Several improvements have been proposed giving rise to MAPSO2 + and MAPSO2 ++. A predictive approach based on PSO called 2s-PSO is also developed to solve the FJSSP problem with consideration of machine failures. Experimental results confirmed that the 2s-PSO algorithm increases the robustness and stability of the solution. As a final contribution, we proposed a rescheduling heuristic to repair schedules in the event of failure under uncertainties.
Complete list of metadatas

Cited literature [98 references]  Display  Hide  Download

https://tel.archives-ouvertes.fr/tel-01640499
Contributor : Maroua Nouiri <>
Submitted on : Monday, November 20, 2017 - 5:51:34 PM
Last modification on : Wednesday, November 22, 2017 - 1:12:48 AM
Long-term archiving on : Wednesday, February 21, 2018 - 2:41:27 PM

File

Thèse_Maroua_Nouiri_Version_F...
Files produced by the author(s)

Identifiers

  • HAL Id : tel-01640499, version 1

Collections

Citation

Maroua Nouiri. Implémentation d'une méta-heuristique embarquée pour résoudre le problème d'ordonnancement dans un atelier flexible de production. Informatique [cs]. Ecole Polytechnique de Tunisie, 2017. Français. ⟨tel-01640499⟩

Share

Metrics

Record views

238

Files downloads

892