Optimization of digital protection strategies for intangible cultural heritage driven by reinforcement learning

China’s intangible cultural heritage (ICH) is made up of things like crafts, music, festivals, and oral traditions which have existed for thousands of years and represent a country’s history and culture. There is now an increasing urgency to protect, revive and adapt ICH to ensure sustainable intergenerational transmission of knowledge due to the acceleration in the rate at which digital technology is transforming our lives. Digital Preservation methods currently in place often do not provide for dynamic adaptability, engagement with the user, and strategic optimization, all of which are necessary for ensuring the long-term success of ICH. To address these challenges, this research proposes a Weighted Bonobo Optimizer-driven Dynamic Deep Q-Network (WBO-DDQ-Net), a hybrid reinforcement learning (RL) model designed to optimize digital protection strategies for ICH. Data were collected from publicly available open-source platforms, which include digital cultural archives, social media analytics (e.g., WeChat, Douyin), and structured user feedback from online surveys. It provides ICH-specific attributes such as category, region, transmission method, engagement metrics, and user sentiment. Two preprocessing techniques, text normalization to standardize dialectal and textual variations and feature scaling to align numeric platform data, were employed to prepare inputs for model training. The model combines the global search capabilities of the WBO with the adaptive policy learning of a DDQ-Net, enabling responsive and intelligent decision-making in heritage dissemination. The WBO-DDQ-Net was trained to maximize a multi-factor reward function including user engagement, cultural continuity, and content diversity. Results show improvement in digital protection using the Python tool, the proposed WBO-DDQ-Net outperformed existing approaches when evaluating accuracy (99.45%). It improves the model’s ability to determine optimal safeguarding strategies with greater accuracy and consistency, outperforming the existing methods across all evaluation metrics. This approach offers a scalable, intelligent framework for modernizing ICH preservation in China’s rapidly evolving digital landscape.