Effects of the Particle Swarm Optimization parameters for structural dynamic monitoring of cantilever beam

Abstract : Nowadays, particle swarm optimization (PSO) algorithm has become a widespread optimization method. However, it is well known that its main parameters (inertia weight, two learning factors, velocity constraint and population size) have a critical effect on its performance. Currently the effects of PSO parameters on structural health monitoring have not been comprehensively studied. Therefore, in this paper, the PSO algorithm is used for damage detection assessment of a cantilever beam, and the simulation results are used to analyze the effects of PSO parameters. There are five levels for each parameter in our experiment, mean fitness value and success rate for each level are used as criteria to measure the convergence and stability of the PSO algorithm. Considering the effect of population size on CPU time, a trade-off strategy is presented to further determine the selection of population size.
Complete list of metadatas

Cited literature [4 references]  Display  Hide  Download

https://hal.archives-ouvertes.fr/hal-02188562
Contributor : Ccsd Sciencesconf.Org <>
Submitted on : Thursday, July 18, 2019 - 3:44:37 PM
Last modification on : Tuesday, November 19, 2019 - 4:48:24 PM

Identifiers

  • HAL Id : hal-02188562, version 1

Citation

Xiao-Lin Li, Roger Serra, Olivier Julien. Effects of the Particle Swarm Optimization parameters for structural dynamic monitoring of cantilever beam. Surveillance, Vishno and AVE conferences, INSA-Lyon, Université de Lyon, Jul 2019, Lyon, France. ⟨hal-02188562⟩

Share

Metrics

Record views

74

Files downloads

49