National Research Council of Canada. Information and Communication Technologies
EUROXXVII Annual Conference, 12-15 July 2015, University of Strathclyde, Glasgow, United Kingdom
OR in distance learning; OR in education; mathematical programming
This work presents the contribution of operational research to education and more particularly to learning design with the implementation of a learning path recommendation system for the next generation of e-learning services. A learning design recommendation system would help learners get appropriate learning objects through an efficient learning path during their self-directed learning journey. The quantity of learning objects available is constantly growing, and millions are now available online. Therefore designing a learning path can be a tedious task that could be eased with the help of software capacities. Moreover, most of the existing recommender solutions proposed by different research communities including educational data mining are not suitable for the very large repositories of learning objects and does not take into account the complexity of the problem in their optimization process. To alleviate this difficulty, we proposed a general approach based on graph theory and mathematical programming to optimize the learning path discovery. The first step of the approach consists in reducing the search space by iteratively building sub-graphs as a succession of cliques form the targeting competencies to competencies reachable by the learner. In a second step, our mathematical model takes into account the prerequisite and gained competencies as constraints and the total competencies needed to reach the learning goal as the objective function to optimize.