实验技术与管理2026，Vol.43Issue(2)：212-218,7.DOI:10.16791/j.cnki.sjg.2026.02.025
基于大语言模型的"操作系统原理"实验教学辅助系统
A large language model-based experimental teaching assistant system for the Principles of Operating Systems course
摘要
Abstract
[Objective]Aiming at the phenomena of overly complicated operation instructions,difficult-to-locate errors,and low problem-solving efficiency in traditional experimental teaching of the Principles of Operating Systems course,this paper designs a"topic task"multi-level experimental teaching method and develops an experimental teaching assistant system based on a large language model.[Methods]For experimental teaching in the Principles of Operating Systems course,a multi-level course system is constructed,including experimental topics,tasks,procedural steps,and unit tests.To provide data support,a fine-tuning dataset of experimental teaching instructions was constructed using large language model-based automatic text generation,and a prompt template was designed to enable the system to comprehensively support experimental operation,verification,and other assisted teaching scenarios.Low-Rank Adaptation(LoRA)-based fine-tuning is applied to the Llama3 large language model,using the constructed dataset and prompt template to enhance its professional question-answering capability for experimental teaching in the Principles of Operating Systems course.[Results]This"topic task"multi-level experimental teaching method helps students clarify experimental objectives,experimental tasks,and procedural steps,and complete tasks and locate errors promptly through unit test-based validation suggestions.Real-world interactive examples demonstrate that the system not only provides students with personalized intelligent Q&A but also automatically generates unit test suggestions for intermediate steps based on experimental tasks.The system decomposes complex experimental operation problems into several simple steps,gradually guiding students through the process of addressing and completing experimental operations and verification.By interacting with the system,students not only receive accurate instructions for experimental operations but can also instantly determine whether each step is correct with automatically generated verification commands,enabling self-checks and self-corrections.[Conclusions]The experimental teaching assistant system for the Principles of Operating Systems course enhances the verifiability of experimental steps and significantly improves both error-checking efficiency and experiment completion rates.The design and implementation of this system represent an effective practice in large language model-assisted computer experiment education,indicating progress in computer experimental teaching.The system also provides scalable research ideas for other computer experiment teaching contexts and is worthy of broader application and promotion across related courses.关键词
实验教学辅助系统/"操作系统原理"课程/"主题—任务"式实训法/Llama3大语言模型/Lora微调Key words
experimental teaching assistant system/Principles of Operating System course/"topic—task"practical training method/Llama3 large language model/Lora-based fine-tuning分类
信息技术与安全科学引用本文复制引用
陶飞飞,张优优,邓劲柏,邹荣辉..基于大语言模型的"操作系统原理"实验教学辅助系统[J].实验技术与管理,2026,43(2):212-218,7.基金项目
教育部产学合作协同育人项目(220600627280456) （220600627280456）
教育部供需对接就业育人项目(20230104537) （20230104537）
江苏省高等教育"电子信息类专业建设、课程建设、教学研究"重点课题(2024JSDZJG12) （2024JSDZJG12）
河海大学本科智慧课程项目(2025-45) （2025-45）
