eprintid: 138064 rev_number: 23 eprint_status: archive userid: 217702 dir: disk0/00/13/80/64 datestamp: 2025-09-08 06:15:41 lastmod: 2025-09-08 06:15:41 status_changed: 2025-09-08 06:15:41 type: thesis metadata_visibility: show creators_name: Ade Mulyana, - creators_name: Yudi Wibisono, - creators_name: Ani Anisyah, - creators_nim: NIM2108799 creators_nim: NIDN0007077506 creators_nim: NIDN0011089304 creators_id: adem01@upi.edu creators_id: yudi@upi.edu creators_id: anianisyah@upi.edu contributors_type: http://www.loc.gov/loc.terms/relators/THS contributors_type: http://www.loc.gov/loc.terms/relators/THS contributors_name: Yudi Wibisono, - contributors_name: Ani Anisyah, - contributors_nidn: NIDN0007077506 contributors_nidn: NIDN0011089304 contributors_id: yudi@upi.edu contributors_id: anianisyah@upi.edu title: NON-PLAYABLE CHARACTER BERBASIS LARGE LANGUAGE MODEL UNTUK GAME ROLE PLAYING GAME ispublished: pub subjects: L1 subjects: Q1 subjects: QA75 divisions: ILKOM full_text_status: restricted keywords: Karakter Non-Pemain, Model Bahasa Besar, Retrieval Augmented Generation, Role-Playing Game, Rekayasa Prompt Non-Playable Characters, Large Language Models, Retrieval Augmented Generation, Role-Playing Game, prompt engineering note: https://scholar.google.com/citations?hl=en&user=mcbS4MAAAAAJ ID SINTA Dosen Pembimbing: Yudi Wibisono: 260167 Ani Anisyah: 6786982 abstract: Game RPG sering menghadirkan dialog NPC yang statis dan kurang kontekstual, sehingga mengurangi imersi pemain terhadap alur cerita. Penelitian ini mengembangkan sistem dialog NPC dinamis berbasis Large Language Model (LLM) dengan pendekatan Retrieval-Augmented Generation (RAG) dan prompt engineering, diimplementasikan pada prototipe game Kage no Meiyaku: Shinobi no Michi. Sistem dirancang agar NPC merespons input pemain secara adaptif dengan merujuk pada lore cerita, riwayat percakapan serta menjaga konsistensi persona karakter. Evaluasi menggunakan kuesioner skala Likert mencakup lima aspek kualitas dialog: relevansi cerita, koherensi naratif, konsistensi persona, kealamian bahasa, dan keterlibatan emosional. Hasil pengujian terhadap 13 responden menunjukkan skor rata-rata 4,07–4,46, menandakan respons positif. Meskipun masih terdapat keterbatasan, seperti kesalahan ejaan pada response, dan fitur game sederhana, sistem terbukti meningkatkan kualitas interaksi dalam game naratif. Penelitian ini tidak hanya menghasilkan prototipe NPC yang lebih interaktif, tetapi juga mengeksplorasi potensi teknologi LLM dalam pengembangan game, khususnya NPC berbasis percakapan yang lebih kontekstual dan imersif. Role-Playing Games (RPGs) often feature static and less contextual NPC dialogues, which can reduce player immersion in the storyline. This study develops a dynamic NPC dialogue system based on Large Language Models (LLMs) using a Retrieval-Augmented Generation (RAG) approach and prompt engineering, implemented in the prototype game Kage no Meiyaku: Shinobi no Michi. The system is designed to enable NPCs to respond adaptively to player input by referencing game lore, conversation history, and maintaining consistent character personas. Evaluation was conducted using a Likert-scale questionnaire covering five dialogue quality aspects: story relevance, narrative coherence, persona consistency, language naturalness, and emotional engagement. Testing with 13 participants showed positive responses, with average scores ranging from 4.07 to 4.46. Despite limitations such as typos in NPC responses and the simplicity of the prototype features, the system proved to enhance interaction quality in narrative-driven games. This research not only delivers a more interactive NPC prototype but also explores the potential of LLM technology in game development, particularly for creating more contextual and immersive conversational NPCs. date: 2025-08-28 date_type: published institution: Universitas Pendidikan Indonesia department: KODEPRODI55201#Ilmu Komputer_S1 thesis_type: other thesis_name: other official_url: https://repository.upi.edu related_url_url: https://repository.upi.edu related_url_type: org citation: Ade Mulyana, - and Yudi Wibisono, - and Ani Anisyah, - (2025) NON-PLAYABLE CHARACTER BERBASIS LARGE LANGUAGE MODEL UNTUK GAME ROLE PLAYING GAME. S1 thesis, Universitas Pendidikan Indonesia. document_url: http://repository.upi.edu/138064/1/S_KOM_2108799_Title.pdf document_url: http://repository.upi.edu/138064/2/S_KOM_2108799_Chapter1.pdf document_url: http://repository.upi.edu/138064/3/S_KOM_2108799_Chapter2.pdf document_url: http://repository.upi.edu/138064/4/S_KOM_2108799_Chapter3.pdf document_url: http://repository.upi.edu/138064/5/S_KOM_2108799_Chapter4.pdf document_url: http://repository.upi.edu/138064/6/S_KOM_2108799_Chapter5.pdf document_url: http://repository.upi.edu/138064/7/S_KOM_2108799_Appendix.pdf