1 Copyright © 2011 by ASME 1 Corresponding Author ASSESSING THE EFFECTIVENESS OF USING GRAVEYARD DATA FOR GENERATING DESIGN ALTERNATIVES Garrett Foster Graduate Research Assistant North Carolina State University gdfoster@ncsu.edu Scott Ferguson 1 Assistant Professor North Carolina State University scott_ferguson@ncsu.edu (919)515-5231 ABSTRACT Modeling to Generate Alternatives (MGA) is a technique used to identify variant designs that maximize design space distance from an initial point while satisfying performance loss constraints. Recent work has explored the application of this technique to nonlinear design problems, where the design space was investigated using an exhaustive sampling procedure. While computational cost concerns were noted, the main focus was determining how scaling and distance metric selection influenced alternative discovery. To increase the viability of MGA for engineering design problems, this work looks to reduce the computational overhead needed to identify design alternatives. This paper investigates and quantifies the effectiveness of using previously sampled designs, i.e. a graveyard, from a multiobjective genetic algorithm as a means of reducing computational expense. Computational savings and the expected error are quantified to assess the effectiveness of this approach. These results are compared to other more common “search” techniques; namely Latin hypercube samplings, grid search, and the Nelder-Mead simplex method. The performance of these “search” techniques are subsequently explored in two case study problems - the design of a two bar truss, and an I-beam - to find the most unique alternative design over a range of different thresholds. Results from this work show the graveyard can be used as a way of inexpensively generating alternatives that are close to ideal, especially nearer to the starting design. Additionally, this paper demonstrates that graveyard information can be used to increase the performance of the Nelder-Mead simplex method when searching for alternative designs. 1. INTRODUCTION Engineering models are full of uncertainty. Designers commonly must ask if the correct assumptions are being made. They must explore if important problem factors were initially ignored, and question the accuracy of the estimated operating conditions. Within this realm of uncertainty designers are asked to arrive at a solution that best solves the existing model of the system. However, this model may not be correct. Working with an invalid or uncertain system model can likely yield non-optimal, or infeasible, designs. A useful tool for a designer includes the ability to create multiple unique solutions that possess similar performance characteristics. This allows a designer to better protect themselves against model uncertainty while simultaneously gaining valuable insight into the problem. A technique developed for this purpose is Modeling to Generate Alternatives (MGA) [1-6]. In previous efforts, the authors have expanded the exploration of MGA to non-linear engineering design problems [6]. This work investigated how the scaling of design variables and the choice of distance metric influenced the generated design alternative. However, this work relied heavily on extensively sampling the design space using a grid search to identify the “best” alternative designs. While this is realizable for simple engineering design problems, the computational overhead quickly becomes overwhelming with increased product architecture. Overcoming this computational barrier is the focus of this paper - namely how can the computational efficiency of MGA be increased such that it is a viable tool for complex engineering design problems? Consider that during the optimization of an engineering design problem a large percentage of the evaluated designs are typically discarded. These evaluated designs comprise what this paper will refer to as a „graveyard‟. It is important to note that a design does not have to be discarded to become part of the graveyard; it merely needs to be evaluated during the optimization routine. In this work, the graveyard is defined as a population of all previously evaluated feasible design which is created during the process of solving a multiobjective optimization problem. Proceedings of the ASME 2011 International Design Engineering Technical Conferences & Computers and Information in Engineering Conference IDETC/CIE 2011 August 28-31, 2011, Washington, DC, USA DETC2011-4