Development of A Multi Objective Search Based System For Solving of Design Optimization Problem


The facing of the difficulties due to high product costs at competition forced industrial institutions to produce low-priced, high quality and reliable products in a short time and various methods are developed as a solution. Although these developed methods were not able to meet the needs at first, insufficiency of these techniques when it is met with complex structure problems, made it obligatory to develop new methods. Since 1980 the researcher have been using living and eating and reproduction processes of living to reach to optimum results. The second stage in developing processes of optimization is methods which use natural process as genetic algorithms, to reach optimum results. The last stage in optimization process is hybrid methods which are integrade two or more methods. In this study, a new hybrid method is proposed to solve single and multi objective test problems and to design vehicle elements which have optimum properties. The proposed new hybrid method based on genetic algorithms and Taguchi's robust design approach. According to the proposed method, It was thought that optimum results could be achieved by the use of levels and level intervals of the design parameters that were obtained as a result of developted single response taguchi approach while forming starting population in genetic algorithms. Since the results which were found with proposed method are better than the results which were found in literature the proposed method can use to solve single and multi-objective optimization problems.