Abstract

Context: Traditional building and infrastructure design methodologies are inflexible and inefficient, leading to high costs and environmental damage. Generative design, with an algorithm that provides multiple options, could be a potential solution. The challenge is creating an intuitive, user-friendly application that optimizes engineers' time, reducing manual iterations and lead to a good user experience (UX). A method for evaluating the UX is heuristic evaluation, in which heuristics are used to inspect a software product. Objective: Since generative design applications have specific features, generic heuristics may not detect all problems related to UX. This article presents a novel set of 9 heuristics to evaluate UX in generative design applications: UXH-GEDAPP. This set is focused on evaluating both UX attributes and specific features of generative design applications. Method: A formal methodology was used to develop the heuristics, through 7 stages: exploratory, descriptive, correlational, selection, specification, validation, and refinement. We performed 3 iterations and validated UXHGEDAPP in 2 iterations through: heuristic evaluation, expert judgment, and user test. Since the methodology can be applied iteratively, we validated and refined the set to improve the proposal. Results: The results obtained in the validation stage indicate that UXH-GEDAPP is useful and more effective than generic heuristics when evaluating generative design applications. UXH-GEDAPP allows to detect specific usability/UX problems as well as more severe problems related to generative design applications. Furthermore, evaluators made fewer errors associating the detected problems with the proposed heuristics, compared to generic sets. Conclusion: UXH-GEDAPP is a new set of heuristics that encourages the creation and use of generative design applications with good UX. It can detect usability/UX problems and help correct them, as well as guide the development of new generative design applications for a pleasant and intuitive user experience.