The incorporation of collaborative work in the educational field grows day after day, as does the research associated with this topic. One of the most recurring problems faced by teachers who want to employ this learning strategy is the good students group formation since this task can be complex both conceptually and computationally, especially if the aim is to automate it. Considering that group formation is key when developing activities in collaborative learning scenarios, this paper aims to propose a technique based on a genetic algorithm approach for achieving homogeneous groups, considering the students personality traits as grouping criteria, since, the aforementioned traits, have proven to be significantly predictive of academic performance and, in addition, to be associated with variables with a strong influence on academic success, such as intelligence and beliefs of self-efficacy. A controlled experiment was designed with 132 students, quantifying the personality traits through the Big Five InventoryBFI, forming working groups and developing a collaborative activity in the initial programming courses. The experiment results allowed validation, not only from a computational point of view evaluating the algorithm performance but also from a pedagogical point of view, confronting the results obtained by students: formation based on personality traits versus formation by the students preference. The highlight of the study is that the homogeneous groups generated by the proposed approach produced better academic results compared to those obtained by groups generated by the students' preference, at the moment of developing a collaborative activity.

This paper presents the analysis of recent research on mobile learning and usability areas, applying a systematic mapping study. The aim is to understand the tendencies and needs in the m-learning field. The results demonstrate that research in the area has grown significantly since 2013, and we identify a necessity when we see that not all the m-learning applications have used usability tests; we also did not find guidelines or frameworks to evaluate them. With these results and tendencies, we propose an evaluation framework for m-learning applications, considering pedagogical usability and user interface usability, to improve the quality of m-learning applications.