Automate the decision on best-suited UI design for mobile apps

Abstract

Recent advances in the mobile domain, such as the introduction of multi-touch gestures, introduce new challenges for interaction designers in producing the best-suited user interface (UI) design and interaction schema. Previously, we have proposed an approach for evolving the created mobile app UIs towards the best-suited design and interaction schema. The approach uses Genetic Algorithm (GA) for searching the best solution (i.e., the possible best UI design and interaction schema). However, the approach works best when given a high number of input populations (i.e., input UI designs), which requires more designers. In this work, we propose to automate the process of generating the candidates input UIs. For this, we propose to use MobiGolog, a formal task modelling language for the mobile domain, for formally specifying the required functionality set alongside the possible UI elements and interaction schema. This formal specification is then used to generate all the possible combinations of UI design layouts and interaction schemas, which are then given as input to the GA in order to give a better solution at the end in a cost effective manner.