Skip to Main content Skip to Navigation

Variations sur PSO : approches parallèles, jeux de voisinages et applications

Abstract : Known for many years as a stochastic metaheuristic effective in the resolution of difficult optimization problems, the Particle Swarm Optimization (PSO) method, however, shows some drawbacks, the most studied: high running time and premature convergence. In this thesis we consider some variants of the PSO method to escape these two disadvantages. These variants combine two approaches: the parallelization of the calculation and the organization of appropriate neighborhoods for the particles. To prove the performance of the proposed models, we performed an experiment on a series of test functions. By analyzing the obtained experimental results, we observe that the proposed models based on the PSO algorithm performed much better than basic PSO in terms of computing time and solution quality. A model based on the PSO algorithm was selected and developed for an experiment on the problem of electricity transmission. A hybrid variant of this model with Simulated Annealing (SA) algorithm has been considered and tested on the problem of collaborative networks.
Complete list of metadatas

Cited literature [81 references]  Display  Hide  Download
Contributor : Abes Star :  Contact
Submitted on : Thursday, August 27, 2020 - 2:35:11 PM
Last modification on : Sunday, September 27, 2020 - 5:10:55 AM


Version validated by the jury (STAR)


  • HAL Id : tel-02923808, version 1


Maria Zemzami. Variations sur PSO : approches parallèles, jeux de voisinages et applications. Complexité [cs.CC]. Normandie Université; Ecole nationale des sciences appliquées (Kénitra, Maroc), 2019. Français. ⟨NNT : 2019NORMIR11⟩. ⟨tel-02923808⟩



Record views


Files downloads