Application of genetic algorithm and Taguchi method in dynamic robust parameter design for unknown problems

Abstract

Taguchi method was known in many industries as an off-line quality control methodology to improve the performance of products or processes at low cost. Although it was effective at improving quality, the statisticians pointed to inefficiencies in the method for highly nonlinear problems and complexity of the product design, the results were often less than satisfactory. Therefore, recently mathematical model and computer simulation are the other alternatives for robust design. However, Taguchi method and computer simulation have their own advantages and disadvantages. So, this idea motivates the approach of combining both advantages together to promote the more effectiveness of robust design. This paper proposes the approach to apply computer-aided engineering (CAE) with genetic algorithm (GA) and Taguchi method in dynamic robust parameter design. Firstly, we use computer simulation, CAE, to obtain the data instead of conducting whole real experiment. Next, a statistical approach, linear regression is used to model the unknown functions, and then GA, heuristics search approach, is employed to find the appropriate setting of controllable factors on the basis of the quality loss function. The effect of noise factors which is the important philosophy of robust design is also considered by utilizing the outer orthogonal array. The objective is to minimize the average quality loss instead of maximizing SN ratio. In addition, to enhance the capability of the proposed approach, the two-step method is applied to screen out the dispersion factors significantly affecting the quality variation and the adjustment factors significantly affecting the sensitivity of the Taguchi’s dynamic system for the case of changing products or processes’ requirements in the future. This proposed is applied to the example “Foam design of automobile’s dashboard”.