Optimization methods

 

 

Date

05/11/2007

Author

F. Moussouni, S. Brisset, P. Brochet

Affiliation

L2EP EC Lille France

Email

fouzia.moussouni@ec-lille.fr, stephane.brisset@ec-lille.fr

Method

Hybrid method

References

[1] F. Moussouni, S. Brisset, P. Brochet, “Some results on design of brushless DC wheel motor using SQP and GA”, International Journal of Applied Electromagnetics and Mechanics I.J.A.E.M. Issue: vol. 26, Num 3-4/2007, P. 233-241

Description of the method

While GA give a well approximation of the global solution, SQP method obtains a more accurate solution in a smaller number of objective function evaluations [2]. Furthermore, if the initial point is far from the global solution, the SQP algorithm can be trapped in local optima. Unlike GA algorithms, which always find the global solution at the first run. While GA allow assuring the global optimum, SQP is an effective local method which quickly finds with accuracy the optimum in the vicinity of the used initial point. Therefore, SQP can be combined with GA to get a new algorihtm called Hybrid method. This one can find the global optimum with a high accuracy and also with a reasonable number of evaluations.

  

Back