A SIMULATION-BASED GENERIC TOOL FOR MULTIDISCIPLINARY DESIGN OPTIMIZATION

INTRODUCTION This paper describes the development of a generic multidisciplinary design optimization tool, which enables any user to mount his or her application program on top of the tool to perform an optimized system design without having to do all the required work in performing multidisciplinary design optimization. This optimum design tool uses the simulation-based approach, in lieu of the traditional search-based approach. The generic tool is written in Microsoft Excel, which combines three different techniques: Taguchi techniques, fuzzy logic and neural networks, into one platform. This tool is designed to be user-friendly and interactive. It is also flexible enough to allow the user to easily switch between single and multiple objective optimizations, and between local and global optimizations, at any time. Furthermore, the user can place a preference weight on each objective. Two types of optimization programs are available to the user. The first one is for the case when the user does not have a database containing input-output pairs. The second one assumes that the user already has such database, and as soon as an optimum design is found, the user has a choice to fine-tune it by slightly changing some design parameter values. * Professor, Mechanical Engineering Department f Research Scientist Copyright © 2000 by the American Institute of Aeronautics and Astronautics, Inc. No copyright is asserted in the United States under Title 17, U.S. code. The U.S. Government has a royaltyfree license to exercise all rights under the copyright claimed herein for Governmental purposes. All other rights are reserved by the copyright owner. The emerging field of Multidisciplinary Design Optimization (MDO) seeks to improve design methodology to rapidly and efficiently explore multipledimension design spaces with the goal of increasing system performance significantly, thereby reducing endproduct cost substantially. Search-based and simulation-based are the two major system design approaches. The former is traditional and mathematical, and has been existing for a long time. The optimum solution has to do with the selected starting point, and the optimization method used. A possibility of divergence in solution seeking is a major drawback in this approach. In contrast, the simulationbased approach uses the analysis and evaluation of a candidate solution, and the assessment of the degree to which the candidate satisfies the requirement. This optimum design tool uses the simulation-based approach'. With this new approach, the optimum solution can be obtained in real time. This powerful tool combines Taguchi techniques, fuzzy logic and neural networks in one platform. Traditionally, global optimization is much more difficult and time consuming to perform using the search-based approach. A common searching technique in global optimization is known as multi-start, in which the user keeps changing the starting point until no better solution can be found. In contrast, the presented design American Institute of Aeronautics and Astronautics (c)2000 American Institute of Aeronautics & Astronautics or Published with Permission of Author(s) and/or Author(s)' Sponsoring Organization. tool can easily perform global optimization as well as local optimization. In local optimization, the user can specify the allowable range from the design baseline. With the simulation-based approach, global optimization is merely a special case of local optimization. COMBINING THE THREE TECHNIQUES FOR OPTIMIZATION In this work, the simulation-based optimization is accomplished by combining Taguchi techniques, fuzzy logic and neural networks. They are briefly described below (A) Taguchi Techniques The Taguchi techniques developed by Dr. Taguchi are cost driven and highly efficient in their ability to extract relatively large amount of information from small experiments'. Taguchi has created a transformation of the repetition data to another value, which is a measure of the variation present. This transformation is the signal-to-noise (S/N) ratio. By examining the S/N ratios, the relative parameter significance can be determined, which provides a basis for parameter settings in optimization later. Table I Taguchi's Orthogonal Array L27 (3) The basic tools used to obtain the information are orthogonal arrays and linear graphs. Three valuable pieces of information are generated from a Taguchi analysis: 1. Which factors or parameters are significant to the output (or objective functions) 2. The relative significance of those factors 3. Which direction for levels of those factors will lead to further improvement or optimization to the design An orthogonal array contains the number of experimental runs, the number of levels of each input parameter, and the number of columns in the array. In an orthogonal array, every input parameter is placed in one of the columns. Table I shows an example of a Taguchi's orthogonal array, L2?(3), which can be used for up to 13 parameters with 3 levels (such as 1 for low, 2 for medium, and 3 for high). In this case, only a total of 27 runs or experiments are needed to identify the significant parameters. (i| 1 1 1 1 1 1 1 1 1 1 1 ff 1 1 1 1 2 2 2 2 2 2 2 111 1 1 1 1 3 3 3 3 3 3 3 1 2 2 2 1 1 1 2 2 2 3 1 2 2 2 2 2 2 3 3 3 1 1 2 2 2 3 3 3 1 1 1 2 i p 1 3 3 3 1 1 1 3 3 3 2 ltj!f 1 3 3 3 2 2 2 1 1 1 3 ( • 1 3 3 3 3 3 3 2 2 2 1 2 1 2 3 1 2 3 1 2 3 1 |fl< 2 1 2 3 2 3 1 2 3 1 2 illl 2 1 2 3 3 1 2 3 1 2 3 2 2 3 1 1 2 3 2 3 1 3 •Jl 2 2 3 1 2 3 1 3 1 2 1 <1|| 2 2 3 1 3 1 2 1 2 3 2 • 2 3 1 2 1 2 3 3 1 2 2 2 3 1 2 2 3 1 1 2 3 3 ; 2 3 1 2 3 1 2 2 3 1 1 : 3 1 3 2 1 3 2 1 3 2 1 3 1 3 2 2 1 3 2 1 3 2 3 1 3 2 3 2 1 3 2 1 3 3 2 1 3 1 3 2 2 1 3 3 3 2 1 3 2 1 3 3 2 1 1 111! 3 2 1 3 3 2 1 1 3 2 2 ll|l 3 3 2 1 1 3 2 3 2 1 2 I|ji 3 3 2 1 2 1 3 1 3 2 3 3 3 2 1 3 2 1 2 1 3 1 2 2 3 3 1 1