Studi Perbandingan Algoritma Finite State Machine, Behavior Tree, Rule-Based System, Dan Utility-Based AI Pada Perilaku NPC Dalam Game Edukasi Fisika Berbasis Gamifikasi

Abstract

This study aims to analyze and compare of four decision-making algorithms for Non-Player Character (NPC) behavior, namely Finite State Machine (FSM), Behavior Tree (BT), Rule-Based System, and Utility-Based AI, within the context of a physics educational game based on gamification. The background of this research is driven by the important role of NPCs in creating engaging interactions that support the learning process, especially in educational games that integrate physics concepts with gameplay mechanics. The implementation of gamification elements such as points, levels, challenges, and interactive feedback is expected to enhance player engagement and motivation in understanding physics concepts. The research method employs an experimental approach by implementing the four algorithms in NPC behavior within a roguelike-based physics educational game. Each algorithm is evaluated based on several indicators, including NPC responsiveness, adaptability to game conditions, implementation complexity, and its impact on player experience. Data collection is conducted through system testing and user questionnaires to measure player engagement and perceptions of NPC behavior. The results indicate that Behavior Tree and Utility-Based AI tend to produce more adaptive and dynamic NPC behaviors compared to Finite State Machine and Rule-Based System. However, Finite State Machine and Rule-Based System demonstrate advantages in terms of simplicity and computational efficiency. From a gamification perspective, algorithms that generate more adaptive NPC behavior significantly improve player engagement. In conclusion, the selection of NPC behavior algorithms greatly influences the quality of interaction in gamified physics educational games. Therefore, developers are advised to consider the balance between algorithm complexity and gameplay requirements when designing effective and educational NPC systems. This study is expected to serve as a reference for developing more interactive and adaptive educational games.