Improving E-Learning Planning Engine Performance Farouq H. Hamed, Khalil Ben Mohamed, and Dickson Lukose MIMOS Berhad, Technology Park 57000 Kuala Lumpur, Malaysia {farouq.hamed | khalil.ben | dickson.lukose}@mimos.my Abstract. Learning Plan (LP) constructors is a very critical component in any modern E-learning system. Researcher in this field focuses on the use of Knowledge Space Theory (KST). Existing approach for constructing LPs depend on application of KST by domain experts to manually construct these plans or by assessing the learners by using some mathematical algorithms. Both approaches require involvement of domain expert. For practical use, these two methods are inflexible and consume large amount of computational resources due to the combinatorial nature of the planning problem. In this paper, we propose a semi-automated LP constructor based on KST that does not require involvement of domain expert and the use of learner’s profile for assigning suitable LP to individual learner. Keywords: E-learning, Planning Engine, Knowledge Space Theory, Personalization 1 Introduction E-learning refers to the use of internet or wireless technologies to deliver a broad array of learning solutions, supported learning and teaching [1]. In 2006, 3.5 million learners were participating in E-learning at institutions of higher education in the United States [2]. According to the Sloan Foundation reports [3], there has been an increase of around 12–14 percent per year on average in enrollments for fully online learning over the five years 2004–2009 in the US post-secondary system, compared with an average of approximately 2 per cent increase per year in enrollments overall. Personalized E-learning systems attempt to reconcile several pieces of information about a learner in order to produce a learning experience that is tailored towards their particular needs [4, 5]. Personalization is one of the key areas of research in this field. Many personalization techniques have been identified [6], but the common approach involve the use of the cognitive profile of the learner to provide personalized interactivity in terms of learning content composition and content delivery modality. Personalized assessment is another area of research [7]. Another approach to personalization is for the E-learning system to create individualized Learning Plans for each of the learner [8]. Constructing LPs is a major challenge for E-learning system, because of the combinatorial nature of the process for constructing the plans. One approach to