Research on the Humanization Design of Game NPCs and User Experience Optimization Based on Large Language Models

DOI

10.2991/978-94-6463-823-3_94How to use a DOI?

Keywords

Large Language Model; Retrieval-enhanced Generation; Knowledge Graph; Reinforcement Learning; NPC Humanized Design

Abstract

With the growing demand for immersive gaming experiences, traditional non-player character (NPC) design has struggled to meet contemporary requirements due to rigid interaction patterns and limited feedback mechanisms. This study systematically explores the application potential of three key technologies—retrieval-augmented generation (RAG), knowledge graph-based question-answering enhancement, and reinforcement learning dialogue optimization—in improving NPC humanization and user experience. RAG technology effectively addresses the “hallucination” problem in large language models by dynamically retrieving external knowledge, thereby enhancing response accuracy, but its effectiveness depends heavily on the quality and scale of the underlying knowledge base. Knowledge graph-based methods enhance consistency between NPC dialogue and game lore by structuring information into triples (subject-predicate-object), though constructing and maintaining these graphs is resource-intensive. Reinforcement learning, when combined with knowledge graphs, accelerates strategy convergence but faces challenges in complex or multi-step tasks. Despite these technologies significantly improving NPC interactivity, issues such as real-time performance, generalization limitations, and high computational costs remain unresolved. This paper provides a cost-efficiency analysis framework to guide developers in selecting the most appropriate technologies based on specific game requirements and outlines future directions for creating more lifelike digital characters, balancing technological innovation with practical implementation constraints.

Copyright

© 2025 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Yuan Feng
AU  - Yanshuangfei Miao
AU  - Yijie Zhou
PY  - 2025
DA  - 2025/08/31
TI  - Research on the Humanization Design of Game NPCs and User Experience Optimization Based on Large Language Models
BT  - Proceedings of the 2025 3rd International Conference on Image, Algorithms, and Artificial Intelligence (ICIAAI 2025)
PB  - Atlantis Press
SP  - 958
EP  - 970
SN  - 2352-538X
UR  - https://doi.org/10.2991/978-94-6463-823-3_94
DO  - 10.2991/978-94-6463-823-3_94
ID  - Feng2025
ER  -