Multi-Objective Evolutionary Design of Mold Temperature Control using DACE for Parameter Optimization

Mehnen, J.^{1, a}; Michelitsch, T.^{1, b}; Lasarczyk, C.; Bartz-Beielstein, T.²

1)

Institut für Spanende Fertigung, Universität Dortmund, Baroper Str. 301, 44227 Dortmund

2)

Lehrstuhl Informatik VII, Universität Dortmund, Otto-Hahn-Str. 16, 44227 Dortmund

a) mehnen@isf.de; b) michelitsch@isf.de

Kurzfassung

The design of mold temperature control strategies (MTCS) is a challenging multi-objective optimization task which demands for advanced optimization methods. Evolutionary algorithms (EA) are powerful stochastically driven search techniques. In this paper an EA is applied to a multi-objective problem using aggregation. The performance of the evolutionary search can be improved using systematic parameter adaptation. The DACE technique (design and analysis of computer experiments) is used to find good MOEA (multi-objective evolutionary algorithm) parameter settings to get improved solutions of the MTCS problem. NEEC, an automatic and integrated software package, which is based on the DACE approach, is applied to find the statistically significant and most promising EA parameters using a sequential parameter optimization (SPO) technique.

Schlüsselwörter

Multi-Objective Evolutionary Algorithms (MOEA), Design and Analysis of Computer Experiments (DACE), Mold Temperature Control Systems (MTCS), New Experimentalism for Evolutionary Computation (NEEC)

Veröffentlichung

In: Short Paper Proceedings of the ISEM 2005, 12th Interdisciplinary Electromagnetic, Mechanic & Biomedical Problem, 2005, Bad Gastein, Pfützner, H.; Leiss, E. (Hrsg.), ISBN 3-902105-00-1, S. 464-465