Towards digital game-based learning content with multi-objective reinforcement learning

Abstract

Game-based learning experiences establish immersive, effective and engaging learning activities. The interest in data-driven learning growth approaches, which personalizes learning experience automatically, has grown over the past years through tips, feedback, and problem scenarios. Learning creators are designed in game-based learning environments to modify simulated activities to meet different goals, such as better learning or cooperation with the pupils, which could be complementary or contradictory. This paper proposes a multi-target enrichment framework to induce game-based learning planners who compromise learning and engagement enhancement in game-based learning environments. To induce a key planner from various student knowledge with a game-based learning experience for the higher education students, we analyze a linear multi-political algorithm, Convex Hull Value Iteration. The findings suggest that a multi-target strengthening curriculum creates more effective policies than single-target interventions in combining multiple rewards. A contextual analysis of basic policy and multifunctional preference vectors shows how an MLS affects the selection of learning activity in student game-based learning experiences. Optimal learning and engagement results are accomplished.