Home // International Journal On Advances in Software, volume 5, numbers 1 and 2, 2012 // View article

An Adaptive Computational Intelligence Algorithm for Simulation-driven Optimization Problems

Authors:
Yoel Tenne
Kazhuiro Izui
Shinji Nishiwaki

Keywords: computational intelligence; modelling; classification; model selection

Abstract:
Modern engineering design optimization often evaluates candidate designs with computer simulations. In this setup, there will often exist candidate designs which cause the simulation to fail and would have no objective value assigned to them. This, in turn, can degrade the effectiveness of the design optimization process and lead to a poor final result. To address this issue, this paper proposes a new computational intelligence optimization algorithm which incorporates a classifier into the optimization process. The latter predicts which candidate designs are expected to cause a simulation failure, and its prediction is used to bias the search towards candidate designs for which the simulation is expected to succeed. However, the effectiveness of this approach depends on the classifier being used, but it is typically not known a-priori which classifier best suits the problem being solved. To address this issue, the proposed algorithm employs a statistically rigorous procedure to autonomously select the classifier type, and to adjust the classifier selection procedure with the goal of improving its accuracy. A performance analysis with a simulation-driven design problem demonstrates the effectiveness of the proposed algorithm.

Pages: 131 to 145

Copyright: Copyright (c) to authors, 2012. Used with permission.

Publication date: June 30, 2012

Published in: journal

ISSN: 1942-2628