What question did this study set out to answer?

This research aims to enhance the adaptiveness and personalization of digital teaching methods in tourism management education using deep reinforcement learning.

March 15, 2026Open Access

Optimization of digital teaching methods for tourism management based on deep reinforcement learning

XWXiali WangLuoyang Normal University

Key Points

This research aims to enhance the adaptiveness and personalization of digital teaching methods in tourism management education using deep reinforcement learning.
Utilized a dataset containing student interaction features like attendance, test scores, and participation.
Applied min–max normalization for consistent feature scaling.
Developed the Intelligent Lévy Flight Distribution-driven Double Deep Q-Network for analysis of learning behaviors.
Employed Double DQN to predict optimal teaching interventions and provide targeted feedback.
Achieved accuracy of 98.23% and precision of 98.33% with the ILFD-Double DQN model.
Maintained low prediction errors ranging from 0.032 to 0.057.
Significantly reduced the time needed for personalized guidance over traditional teaching methods.

Abstract

The rapid advancement of digital technologies has transformed tourism management education, creating opportunities to enhance teaching effectiveness and student engagement. However, current digital teaching approaches suffer from low adaptiveness and weak personalization. To overcome these limitations, this research explores the optimization of digital teaching methods in tourism management courses through the application of Deep Reinforcement Learning (DRL). A student interaction dataset is used, containing features such as attendance, test scores, and participation, enabling accurate modeling of learning behaviors. Min–max normalization scales these features, ensuring consistency and effective training for the model. The proposed framework, the Intelligent Lévy Flight Distribution-driven Double Deep Q-Network (ILFD-Double DQN) model, is used to analyze student interactions, learning behaviors, and performance metrics in a digital learning environment. Double DQN is employed to analyze student behaviors, predict optimal teaching interventions, and provide targeted feedback for improved learning outcomes. The Intelligent ILFD uses a Quasi Opposition Based strategy that enhances exploration, enabling dynamic adaptation of content, strategies, and resources for effective, real-time adaptive learning. Experimental evaluation demonstrates that the ILFD-Double DQN outperforms traditional digital teaching approaches, achieving accuracy at 98.23% and precision at 98.33%. The model maintains low prediction errors, ranging between 0.032 and 0.057, while significantly reducing the time required for personalized guidance. These findings highlight the potential of DRL to create a responsive, adaptive, and efficient digital learning environment in tourism management education. It provides actionable insights for educators and institutions seeking to integrate intelligent, data-driven strategies into teaching practices, ultimately enhancing student competence, employability, and readiness for industry demands.

AI에게 질문

Bookmark

View Full Paper