Development of visual design steering as an aid in large-scale multidisciplinary design optimization. Part I: method development
A modified paradigm of Computational Steering (CS), termed Visual Design Steering (VDS), is developed in this paper. The VDS paradigm is applied to optimal design problems to provide a means for capturing and enabling designer insights. VDS allows a designer to make decisions before, during or after an analysis or optimization via a visual environment, in order to effectively steer the solution process. The objective of VDS is to obtain a better solution in less time through the use of designer knowledge and expertise. Using visual representations of complex systems in this manner enables human experience and judgement to be incorporated into the optimal design process at appropriate steps, rather than having traditional black box solvers return solutions from a prescribed input set. Part I of this paper focuses on the research issues pertaining to the Graph Morphing visualization method created to represent an n-dimensional optimization problem using 2-dimensional and 3-dimensional visualizations. Part II investigates the implementation of the VDS paradigm, using the Graph Morphing approach, to improve an optimal design process. Specifically, the following issues are addressed: impact of design variable changes on the optimal design space; identification of possible constraint redundancies; impact of constraint tolerances on the optimal solution; and smoothness of the objective function contours. It is demonstrated that Graph Morphing can effectively reduce the complexity and computational time associated with some optimization problems.
Unable to display preview. Download preview PDF.