Scholarlink

<p>Traditional teaching strategies often fail to dynamically adapt leading to suboptimal student engagement. However, traditional teaching strategies often fail to dynamically adapt to diverse classroom conditions, leading to suboptimal student engagement and uneven participation. This research addresses this problem by proposing an optimization framework for drama teaching strategies known as the Boosted Beluga-optimized Dual-Priority Replay-Driven Dueling Double Deep Q-Network (BBO-D3QN). Classroom data, including student participation rates and feedback ratings, were first collected and normalized during pre-processing to ensure consistency and comparability. Feature extraction using Principal Component Analysis (PCA) was applied to reduce dimensionality and emphasize the most relevant engagement indicators. A simulation environment, implemented in Python 3.10, was constructed to model drama classroom interactions, providing a controlled and repeatable setting for training and evaluation. The proposed approach employs a Dual-Priority Replay-Driven Dueling Double Deep Q-Network (D3QN) to learn optimal teaching strategies by considering both participation and feedback as reward signals. To further enhance convergence speed and policy stability, a Boosted Beluga Whale Optimization (B-BWO) algorithm was integrated for hyperparameter tuning and network optimization. The BBO-D3QN framework demonstrated significant improvements over heuristic and baseline DRL models, with increased student participation, higher average feedback ratings, and improved policy stability. Experimental evaluation confirmed that this integration led to faster convergence and superior cumulative reward performance. Overall, the proposed DRL-based framework provides a robust and adaptive approach to drama teaching strategy optimization, achieving an F1-score of 0.86. By combining advanced reinforcement learning with metaheuristic optimization, this research highlights a pathway toward intelligent, data-driven drama education systems that enhance engagement, creativity, and overall learning outcomes.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Optimization of drama teaching strategies driven by deep reinforcement learning

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月11日

查看全文
  • Zihua Wang

摘要

Traditional teaching strategies often fail to dynamically adapt leading to suboptimal student engagement. However, traditional teaching strategies often fail to dynamically adapt to diverse classroom conditions, leading to suboptimal student engagement and uneven participation. This research addresses this problem by proposing an optimization framework for drama teaching strategies known as the Boosted Beluga-optimized Dual-Priority Replay-Driven Dueling Double Deep Q-Network (BBO-D3QN). Classroom data, including student participation rates and feedback ratings, were first collected and normalized during pre-processing to ensure consistency and comparability. Feature extraction using Principal Component Analysis (PCA) was applied to reduce dimensionality and emphasize the most relevant engagement indicators. A simulation environment, implemented in Python 3.10, was constructed to model drama classroom interactions, providing a controlled and repeatable setting for training and evaluation. The proposed approach employs a Dual-Priority Replay-Driven Dueling Double Deep Q-Network (D3QN) to learn optimal teaching strategies by considering both participation and feedback as reward signals. To further enhance convergence speed and policy stability, a Boosted Beluga Whale Optimization (B-BWO) algorithm was integrated for hyperparameter tuning and network optimization. The BBO-D3QN framework demonstrated significant improvements over heuristic and baseline DRL models, with increased student participation, higher average feedback ratings, and improved policy stability. Experimental evaluation confirmed that this integration led to faster convergence and superior cumulative reward performance. Overall, the proposed DRL-based framework provides a robust and adaptive approach to drama teaching strategy optimization, achieving an F1-score of 0.86. By combining advanced reinforcement learning with metaheuristic optimization, this research highlights a pathway toward intelligent, data-driven drama education systems that enhance engagement, creativity, and overall learning outcomes.