Forming well-balanced collaborative learning teams according to the roles of their members: An evolutionary approach

VIRGINIA YANNIBELLI; ANALÍA AMANDI

Budapest

Simposio; IEEE CINTI 2011 (12th IEEE International Symposium on Computational Intelligence and Informatics); 2011

Óbuda University, Hungary; Hungarian Fuzzy Association; IEEE Computational Intelligence Society; IEEE Systems, Man, and Cybernetics Society; IEEE Industrial Electronics and Robotics and Automation Societies; John von Neumann Computer Society

In university software engineering courses, professors usually divide students into collaborative learning teams to perform collaborative learning tasks (e.g., complete software projects). In this context, one of the grouping criteria most utilized by professors is based on the students’ roles and on forming well-balanced teams according to the roles of their members. However, the implementation of this criterion requires a considerable amount of time, effort and knowledge on the part of the professors. In this paper, we address the problem of forming well-balanced learning teams according to the roles of their members automatically. To solve the problem, we propose an evolutionary algorithm. Considering a given number of students who must be divided into a given number of teams, the algorithm both designs different alternatives to divide students into teams and evaluates each alternative as regards the grouping criterion previously mentioned. This evaluation is carried out on the basis of knowledge of the students’ roles. To evaluate the performance of the algorithm, we report the computational experiments developed on ten data sets with different levels of complexity. The obtained results are really promising since the algorithm has reached high-quality solutions in an acceptable computation time for each of the data sets.