Gary Stump

Design Space Visualization and Its Application to a Design by Shopping Paradigm

We have developed a data visualization interface that facilitates a design by shopping paradigm, allowing a decision-maker to form a preference by viewing a rich set of good designs and use this preference to choose an optimal design. Design automation has allowed us to implement this paradigm, since a large number of designs can be synthesized in a short period of time. The interface allows users to visualize complex design spaces by using multi-dimensional visualization techniques that include customizable glyph plots, parallel coordinates, linked views, brushing, and histograms. As is common with data mining tools, the user can specify upper and lower bounds on the design space variables, assign variables to glyph axes and parallel coordinate plots, and dynamically brush variables. Additionally, preference shading for visualizing a user’s preference structure and algorithms for visualizing the Pareto frontier have been incorporated into the interface to help shape a decision-maker’s preference. Use of the interface is demonstrated using a satellite design example by highlighting different preference structures and resulting Pareto frontiers. The capabilities of the design by shopping interface were driven by real industrial customer needs, and the interface was demonstrated at a spacecraft design conducted by a team at Lockheed Martin, consisting of Mars spacecraft design experts.Copyright

Trade space exploration of satellite datasets using a design by shopping paradigm

One of the goals of early stage conceptual design is to execute broad trade studies of possible design concepts, evaluating them for their capability to meet minimum requirements, and choosing the one that best satisfies the goals of the project. To support trade space exploration, we have developed the advanced trade space visualizer (ATSV) that facilitates a design by shopping paradigm, which allows a decision-maker to form a preference a posteriori and use this preference to select a preferred satellite. Design automation has allowed us to implement this paradigm, since a large number of designs can be synthesized in a short period of time. The ATSV uses multidimensional visualization techniques, preference shading, and Pareto frontier display to visualize satellite trade spaces.

Visual Steering Commands for Trade Space Exploration: User-Guided Sampling With Example

Recent advancements in computing power and speed provide opportunities to revolutionize trade space exploration, particularly for the design of complex systems such as automobiles, aircraft, and spacecraft. In this paper, we introduce three Visual Steering Commands to support trade space exploration and demonstrate their use within a powerful data visualization tool that allows designers to explore multidimensional trade spaces using glyph, 1-D and 2-D histogram, 2-D scatter, scatter matrix, and parallel coordinate plots; linked views; brushing; preference shading and Pareto frontier display. In particular, we define three user-guided samplers that enable designers to explore (1) the entire design space, (2) near a point of interest, or (3) within a region of high preference. We illustrate these three samplers with a vehicle configuration model that evaluates the technical feasibility of new vehicle concepts. Future research is also discussed.Copyright

Visual Steering Commands and Test Problems to Support Research in Trade Space Exploration

Designers can simulate thousands, if not millions, of design alternatives more cheaply and quickly than ever before with today’s computing power; however, the resulting data can overwhelm designers without proper tools to support multi-dimensional data visualization. In this paper, we discuss the use of a multi-dimensional data visualization tool and visual steering commands which allow designers to navigate multi-attribute trade spaces. The novelty in our work is providing designers with a set of visual steering commands to simultaneously explore the trade space and exploit new information and insights as they are gained. Specifically, designers can explore the entire design space (either sampled randomly or manually) or along the entire Pareto front using the Basic Sampler, Point Sampler, and/or Pareto Sampler. Alternatively, they can exploit information they have gained during the exploration process by searching near a specific point of interest or within a region of high preference using the Attractor, Preference Sampler, and/or Guided Pareto Sampler. Examples of each are included in this paper. Meanwhile, a suite of test problems is being formalized to support our trade space exploration – algorithmic development as well as empirical studies involving human decision-makers. This work supports our long-term goal of quantifying the benefits of putting humans back “in-the-loop” during design optimization.

Assessing the Impact of Graphical Design Interfaces on Design Efficiency and Effectiveness

Despite the apparent advantages of and recent advances in the use of visualization in engineering design and optimization, we have found little evidence in the engineering literature that assesses the impact of fast, graphical design interfaces on the efficiency and effectiveness of engineering design decisions or the design optimization process. In this paper we discuss two examples—the design of an I-beam and the design of a desk lamp— for which we have developed graphical and text-based design interfaces to assess the impact of having fast graphical feedback on design efficiency and effectiveness. Design efficiency is measured by recording the completion time for each design task, and design effectiveness is measured by calculating the error between each submitted design and the known optimal design. The impact of graphical feedback is examined by comparing user performance on the graphical and text-based design interfaces while the importance of rapid feedback is investigated by comparing user performance when response delays are introduced within each design interface. Experimental results indicate that users of graphical design interfaces perform better (i.e., have lower error and faster completion time) on average than those using text-based design interfaces, but these differences are not statistically significant. Likewise, we found that a response delay of 0.5 seconds increases error and task completion time, on average, but these increases are not always statistically significant. Trials using longer delays of 1.5 seconds did yield significant increases in task completion time. We also found that the perceived difficulty of the design task and using the graphical interface controls were inversely correlated with design effectiveness—designers who rated the task more difficult to solve or the graphical interface more difficult to use actually performed better than those who rated them easy. Finally, a significant ‘‘playing’’ effect was observed in our experiments: those who played video games more frequently or rated the slider bars and zoom controls easy to use took more time to complete the design tasks. @DOI: 10.1115/1.1583757#

Trade Space Exploration of a Wing Design Problem Using Visual Steering and Multi-Dimensional Data Visualization

Trade space exploration is a promising decision-making paradigm that provides a visual and intuitive means for formulating, adjusting, and ultimately solving multi-objective design optimization problems. This is achieved by combining multi-dimensional data visualization techniques with visual steering commands to allow designers to “steer” the optimization process while searching for the best, or Pareto optimal, designs. In this paper, we investigate the impact of constraint handling on the trade space exploration process. Specifically we consider three different constraint handling methods: (1) no constraint handling, (2) manual constraint handling, and (3) automatic constraint handling, and assess their impact on the efficiency and effectiveness of the visual steering commands used to explore the trade space. We find that the performance of the constraint handling method is highly correlated with the visual steering command that is being used and is consistent with the user’s a priori knowledge about the constraints, which is reflected in how constraints are handled in each method. The implications of these findings on the trade space exploration process are also discussed in conjunction with future work.

Visual Steering and Trade Space Exploration

The assumptions at the beginning of a trade space exploration are that the decision makers have a model of some complex engineered system that relates design variables to performance and cost metrics, they know what the inputs and outputs to the model are, and they know they will be forming a preference over some subset of the inputs and outputs. What they do not know is the relationship between the inputs and outputs (in exact or an intuitive sense), the feasible range of inputs and outputs, the subset of inputs and outputs they will form their preference on, or the exact form of the preference. These assumptions probably form the least informative starting point for trades. To conduct the trade study the model is tied to an exploration engine, which initially randomly exercises the model, creating different system concepts. A user simultaneously visually explores the trade space in real time as it emerges using multi-dimensional data visualization tools and then visually steers further model runs to desired trade space regions of interest by specifying attractors in the trade space, such as desired inputs, outputs, preference functions, Pareto frontier. To ground the presentation, the paper uses a satellite design model, which relates design and performance variables to form a multi-dimensional trade space for satellite configurations. The trade space is discontinuous and complex, and presents a suitable test case.

Preference Construction, Sequential Decision Making, and Trade Space Exploration

This paper develops and explores the interface between two related concepts in design decision making. First, design decision making is a process of simultaneously constructing one’s preferences while satisfying them. Second, design using computational models (e.g., simulation-based design and model-based design) is a sequential process that starts with low fidelity models for initial trades and progresses through models of increasing detail. Thus, decision making during design should be treated as a sequential decision process rather than as a single decision problem. This premise is supported by research from the domains of behavioral economics, psychology, judgment and decision making, neuroeconomics, marketing, and engineering design as reviewed herein. The premise is also substantiated by our own experience in conducting trade studies for numerous customers across engineering domains. The paper surveys the pertinent literature, presents supporting case studies and identifies use cases from our experiences, synthesizes a preliminary model of the sequential process, presents ongoing research in this area, and provides suggestions for future efforts.Copyright

Advanced Visualization Techniques for Trade Space Exploration

Recent advances in modeling and simulation technology have made it feasible to generate large datasets of design alternatives and their attributes in a relatively short amount of time. However, tools to understand and explore these datasets are limited. To this end, the Applied Research Laboratory at Penn State University has been developing a tool, entitled the ARL Trade Space Visualizer (ATSV) to support multi-dimensional trade space exploration. The ARL, in conjunction with the Lockheed Martin Corporation, has extended the tool to tackle several real world design challenges. In response to the needs of the engineering teams at Lockheed Martin, several key enhancements to the ATSV have been designed and implemented. These enhancements include contour plotting in two dimensions; isosurface generation in three dimensions; multiple independent brushing controls; and k-means cluster analysis. This paper will describe the full capabilities of the tool, as well as give an example of the types of design optimization performed by Lockheed Martin. The paper will focus on using the advanced visualization techniques to discover relationships within the dataset that would otherwise prove difficult to extract using traditional analysis techniques.

Interactive Design Selection Process through Visualization and User Guided Search

ठA new process is de scribed that assists engineers in making complex decision s during the design process . The Interactive Design Selection Process (iDSP) allows designers to collect design information , compar e candidate designs , build use r preferences , and ultimately zero in on the best designs. In contrast to traditional black -box optimization approaches, the iDSP involves the designer in each and every step of the process. This involvement gives designers more confidence in the results a nd allows them to find design s that satisfy all stake -holders. A prototype of the iDSP was implemented by combining an advanced visualization tool with a genetic optimizer using a commercially available process integration environment. A Lunar exploration mission design problem was solved using the guided search technique augmented by design space visualization.