THE EFFECTS OF COUPLING ON PARAMETER DESIGN

Axiomatic Design holds that uncoupled designs are to be preferred over coupled designs. This paper documents an effort to empirically quantify the effects of coupling on the design process and how those effects depend on the scale of the problem (number of design variables). An experiment was conducted with human subjects who solved parameter design tasks through a simple graphical user interface. It is established that, for parameter design tasks with only two inputs and two outputs, coupling had only a moderate effect on the subjects’ solution of the problem. As the number of variables increases, the effect of coupling among variables has a drastic effect on the solution procedures and the completion time. The time for a human to solve a coupled parameter design problem rises geometrically as problem size rises from 2X2 to 5X5. These results are discussed in the context of information processing models of human cognition. The implications for Axiomatic Design are discussed