How sensitivity analysis can improve metaheuristic convergence

Peio Loubière; Astrid Jourdan; Patrick Siarry; Rachid Chelouah

Download PDF Open PDF in browser

Switch back to the title and the abstract in French

How sensitivity analysis can improve metaheuristic convergence

EasyChair Preprint no. 2525

2 pages•Date: February 1, 2020

Peio Loubière, Astrid Jourdan, Patrick Siarry and Rachid Chelouah

Abstract

The paper presents the coupling of a sensitivity analysis (SA) method and a metaheuristic (here PSO). The SA method will analyse a certain number of evaluations made by the metaheuristic to compute effect of each input variable on the objective function results (linear, non linear, negligible).

According to these results, for each variable, an indicator is deduced and added in the moving equation. The goal is to increase the shift of a viriable which has a linear effect and reduce it in case of non linear effect.

Though a toy example, we illustrate the behavior and resullts of our method.

Keyphrases: Analyse de sensibilité, métaheuristiques, Optimisation

Links:

https://easychair.org/publications/preprint/pCP4

BibTeX entry

BibTeX does not have the right entry for preprints. This is a hack for producing the correct reference:

@Booklet{EasyChair:2525,
  author = {Peio Loubière and Astrid Jourdan and Patrick Siarry and Rachid Chelouah},
  title = {How sensitivity analysis can improve metaheuristic convergence},
  howpublished = {EasyChair Preprint no. 2525},

  year = {EasyChair, 2020}}

Download PDF Open PDF in browser