Productive iteration in student engineering design projects
Iteration in design has different meanings, ranging from simple task repetition to heuristic reasoning processes. Determining productive iterations is important to improve the design process on cost, time, and quality, but currently there is no categorization of iterations conducive to this goal. After exploring the possible causes and attempts to address them, I propose to classify iterations as rework, design, or behavioral. The framework suggests that design teams, to improve productivity, should try to eliminate rework by increasing the resolution of design information (design iterations) without skipping design levels and by developing alternative solutions (behavioral iterations) in parallel before selecting one. Analysis of journal data from twelve student projects helps identify design processes that achieve higher quality in less time. Factor analysis groups common variability into factors. A multivariate linear regression model of three factors explains 91% of productivity variance within the study sample. Factor scoring coefficients are then used to translate the regression model coefficients back to activities and design levels. Results indicate that generating ideas and defining the problem at a system level are the key discriminating variables between more or less productive design teams in the sample, which supports the recommendation of increasing the resolution of design information without skipping intermediate levels. If we consider selecting an alternative for the final solution as the main design decision students make in the sample projects, then work on non-selected alternatives before selecting the final design can be used as a proxy for effort allocated to behavioral iterations. A linear model using work on non-selected alternatives shows that generating ideas at a system level relates to higher productivity while refining design details and evaluating existing design configurations associate with lower productivity. Then behavioral iteration relates to higher productivity only if alternatives are developed to the system level by generating ideas on how to address interface and configuration issues. The framework presented in this thesis helps differentiate between productive and less productive iteration patterns and provides guidelines to prevent rework by allocating more effort in productive iteration, namely behavioral and design iteration.