100 likes | 183 Views
VisualizeIT: Measuring the Impact of IT-Enabled Concept Generation on Designer Creativity. 101001010010010011100100110111010010111001001001110001010. Rob Stone, Seth Orsborn - Missouri S&T Kemper Lewis, Ken English - U. Buffalo Dan McAdams, Julie Linsey - Texas A&M
E N D
VisualizeIT: Measuring the Impact of IT-Enabled Concept Generation on Designer Creativity 101001010010010011100100110111010010111001001001110001010 • Rob Stone, Seth Orsborn - Missouri S&T • Kemper Lewis, Ken English - U. Buffalo • Dan McAdams, Julie Linsey - Texas A&M • Matt Campbell - U. Texas at Austin
VisualizeIT VisualizeIT 10110 10110 1001 1001 0100 0100 100 100 Research Objective • Starting point: automated concept generation • To measure the effect that different visualization and concept clustering techniques have on designer creativity • Cluster the many concept variants from the automated concept generation algorithms • Represent the option space to the designer so that it enhances creativity • Measure the impact of the visualization schemes on designer creativity
VisualizeIT 10110 1001 0100 100 VisualizeIT
VisualizeIT INPUT: Function Structure 10110 1001 0100 100 3 5 4 OUTPUT: CFG created by execution of rule sequence 3-5-4.
VisualizeIT PCA-based clustering 10110 1001 0100 100
VisualizeIT 10110 1001 0100 100
VisualizeIT Concept Explorer • The VisualizeIT Concept Explorer is being developed to enable the exploration of the large number of possible design candidates. • Each design concept is represented on a virtual card • Components are represented on the top of the card. • Design performance is represented by the radar plot at the lower left.
VisualizeIT Concept Explorer • A user can also use that design to seed the concept search. • For example, the fourth component here was desirable, and a new set of concepts was generated based on that fact. • The interface starts by displaying 8-10 sample concepts from the concept generator. • If a particular concept is of interest, the user can select that card for further investigation. • The performance details can be viewed, as well as more details on how components map to functional requirements. • Users navigate through the concepts using the mouse to spin through the options.
VisualizeIT 10110 1001 0100 100 Experimental Evaluation • Hypothesis: Presentation of computationally generated solutions will produce more complete and innovative concepts. • Groups: Control- no computational support • Experimental- given computationally created concepts • Metrics • Number of ideas • Quality • Completeness • Novelty • Variety
VisualizeIT 10110 1001 0100 100 Design Problem: Flavor Composition Device • Task: Design a device that automatically measures, combines and dispenses spices.