The use of swarm intelligence techniques in e-learning scenarios provides a way to combine simple interactions of individual students to solve a more complex problem. After getting some data from the interactions of the first students with a central system, the use of these techniques converges to a solution that the rest of the students can successfully use. This paper uses a case study to analyze how fast swarm intelligence techniques converge when applied to solve the problem of e-learning resource filtering. Some modifications to traditional ant colony optimization (ACO) algorithms based on student filtering are also introduced in order to improve convergence.