A Novel Approach to Optimization Problem without Objective Function
- https://doi.org/10.2991/iske.2007.272How to use a DOI?
- optimization; modeling; neural network; genetic algorithm
For optimization problem in complex systems, normally, the objective funtion with respect to decision-making variables is hardly obtained or quantified. In this paper, a novel approach is presented. Its first procedure is to model the objective function by fitting experimental data with NN. Secondly, global optimization solution would be acquired for the fitted objective function with genetic algorithm (GA). In order to verify the idea, Peaks function inside MATLAB was selected to demostrate the approach. The fitted Peaks function was gained through training a proper NN with complete input and output data pairs, and then global optimal solution of Peaks function and the fitted Peaks function were searched with genetic algorithm (GA), respectively. Results show that their optimal values and the correspnding solutions were both very close. Furthermore, the modeling and optimization method in the paper was also effectively applied in optimizing the comprehensive performance in analog PID control system. Therefore, the methodology, which combines modeling approach of NN with global optimization of GA, could effectively solve many optimizing puzzles in complex systems without objective function.
- © 2007, the Authors. Published by Atlantis Press.
- Open Access
- This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Cite this article
TY - CONF AU - Yingying Su AU - Wenjin Hu AU - Ling Nie AU - Wen Ye AU - Taifu Li PY - 2007/10 DA - 2007/10 TI - A Novel Approach to Optimization Problem without Objective Function BT - Proceedings of the 2007 International Conference on Intelligent Systems and Knowledge Engineering (ISKE 2007) PB - Atlantis Press SP - 1589 EP - 1593 SN - 1951-6851 UR - https://doi.org/10.2991/iske.2007.272 DO - https://doi.org/10.2991/iske.2007.272 ID - Su2007/10 ER -