Solving the Individualized Instructional Content Delivery Problem Using Ontology and Metaheuristics on the Set Covering Problem: An Experimental Analysis

Abstract

Intelligent Tutoring Systems (ITSs) based on a step-by-step problem-solving approach are limited in terms of compatible content. On the other hand, recommendation systems can suggest various content types but lack the granularity of concepts found in step-by-step approaches. This study addresses this challenge by proposing a method to recommend instructional content from diverse knowledge domains while incorporating the refined concepts of ITSs. To tackle this issue, the instructional content delivery problem (LORP) is formulated as a set covering problem, classified as NP-hard. We show that a PSO-based algorithm is a good candidate to solve LORP in a better runtime than the exact algorithm and with better solutions than the greedy heuristic. By leveraging collaborative filtering and an ontology that models students’ knowledge, learning styles, and search parameters, the approach offers more individualized content.