Reinforcement Learning-Enabled NPC Behavior and Its Influence on Player Emotional States

Non-Playable Characters (NPCs) are central to player immersion in modern digital games. Recent advances in artificial intelligence (AI) have enabled NPCs to move beyond scripted behavior toward adaptive, emotionally responsive, and context-aware agents. This paper presents a structured review and conceptual framework for embedding AI techniques in game NPCs, focusing on pathfinding, behavior and decision trees, emotional modeling, and adaptive difficulty mechanisms. The study synthesizes existing approaches, proposes an integrated NPC behavior architecture, and discusses its influence on player emotional engagement and realism. The rewritten content follows IEEE journal standards and has been paraphrased and reorganized to ensure originality and minimize plagiarism