Hierarchical and interactive decision refinement methodology for engineering design

This paper presents the Hierarchical and Interactive Decision Refinement (HIDER) methodology for engineering design. This methodology hierarchically refines (or reduces) a large initial design space through a series of multiple-objective optimizations, until a fully specified design is obtained. This contrasts with the traditional approach to design optimization which involves picking a fully specified design from the space of possible designs and then iteratively modifying its specifications to maximize the performance. The HIDER methodology uses an adaptive modeling approach that combines machine learning and statistical techniques for developing fast empirical models with different levels of detail and evaluation speeds. These models are used for multiple-objective optimization at different stages of the design. The use of layered models enables HIDER methodology to be used at different stages of parametric design, including early stages when design specifications are not complete. This methodology also supports multiple perspectives in concurrent design because it provides means for transforming models from these different perspectives into a uniform representation, as well as facilitating decision-making with respect to multiple competing objectives. This paper presents the HIDER methodology with an example in the parametric design of a diesel engine.

[1]  John B. Heywood,et al.  Development and Use of a Computer Simulation of the Turbocompounded Diesel System for Engine Performance and Component Heat Transfer Studies , 1986 .

[2]  John R. Dixon,et al.  A review of research in mechanical engineering design. Part II: Representations, analysis, and design for the life cycle , 1989 .

[3]  Mark D. Gross,et al.  Constraints: Knowledge representation in design , 1988 .

[4]  Mokhtar S. Bazaraa,et al.  Nonlinear Programming: Theory and Algorithms , 1993 .

[5]  H. Saunders,et al.  Finite element procedures in engineering analysis , 1982 .

[6]  George E. P. Box,et al.  Empirical Model‐Building and Response Surfaces , 1988 .

[7]  Raymond H. Myers,et al.  Probability and Statistics for Engineers and Scientists. , 1973 .

[8]  Ching-Lai Hwang,et al.  Mathematical programming with multiple objectives: A tutorial , 1980, Comput. Oper. Res..

[9]  Stephen C. Y. Lu,et al.  Creating and using models for engineering design: a machine-learning approach , 1992, IEEE Expert.

[10]  Andrzej Osyczka,et al.  Multicriteria Design Optimization: Procedures and Applications , 1990 .

[11]  John R. Dixon,et al.  A review of research in mechanical engineering design. Part I: Descriptive, prescriptive, and computer-based models of design processes , 1989 .

[12]  S.C-Y. Lu,et al.  Integration of Simulation, Learning and Optimization to Support Engineering Design , 1991 .

[13]  William N. Sharpe,et al.  Evaluation of a modified monotonic Neuber relation , 1991 .

[14]  Larry A. Rendell,et al.  A New Basis for State-Space Learning Systems and a Successful Implementation , 1983, Artif. Intell..

[15]  Garret N. Vanderplaats,et al.  Numerical Optimization Techniques for Engineering Design: With Applications , 1984 .

[16]  R. H. Myers,et al.  Probability and Statistics for Engineers and Scientists , 1978 .

[17]  R. Haftka,et al.  Structural shape optimization — a survey , 1986 .

[18]  S. C.-Y. Lu,et al.  Building Layered Models to Support Engineering Decision Making: A Machine Learning Approach , 1991 .

[19]  Robert B. Haber,et al.  Design Sensitivity Analysis for Elastodynamic Systems , 1990 .

[20]  G. M. Zhang,et al.  Simulation of Surface Topography Formed During the Intermittent Turning Processes , 1991 .

[21]  Susan Finger.  Design Fusion : a product life-cycle view for engineering designs , 1990 .