Grid-based Genetic Operators for Graphical Layout Generation

Graphical user interfaces (GUIs) have gained primacy among the means of interacting with computing systems, thanks to way they leverage human perceptual and motor capabilities. However, design GUIs has mostly been a manual activity. To GUI, designer must select its visual, spatial, textual, interaction properties such that combination strikes balance relevant factors. While emerging computational-design techniques addressed some problems related grid layouts, no general approach proposed can also produce good complete results covering color-related decisions other nonlinear objectives. Evolutionary algorithms are promising demonstrate handling similar in conditions, genetic operators, depending on how designed. But even these approaches struggle elements' overlap hence too many infeasible candidate solutions. This paper presents new based grid-based operators demonstrated non-dominated sorting algorithm (NSGA-III) setting. The use lines for element positions novel manner satisfy overlap-related constraints intrinsically improve alignment elements. be used crossovers mutations. Its core benefit is all solutions generated no-overlap requirement represent well-formed layouts. permit using increasingly realistic task instances, responding more objectives than could considered before. Specifically, we address quality, alignment, selection time, clutter minimization, saliency control, color harmony, grouping