Scholarlink

<p>Under the dual challenges of global warming and environmental degradation, the global demand for energy continues to escalate, accompanied by an increasing emphasis on technological innovation. The new energy sector has emerged as a pivotal driver for sustainable development. This study investigates China’s new energy industry through the integration of complex network theory and deep reinforcement learning models. By synergizing complex network analysis with the Deep Deterministic Policy Gradient (DDPG) algorithm, we quantitatively examine risk interdependencies and node significance among listed enterprises using statistical metrics, ultimately aiming to achieve portfolio risk diversification and return enhancement. A multi-layer complex network model is constructed to analyze risk correlations within the new energy sector. A composite evaluation index is developed to identify low-centrality stocks, forming a stock pool characterized by reduced risk interconnectedness. The DDPG algorithm is then applied to optimize portfolio allocation strategies, balancing risk diversification with the dual objectives of maximizing returns and minimizing volatility. In conclusion, this research demonstrates that complex networks can quantitatively describe the degree of interconnection among assets in the new energy sector. By comparing low-centrality portfolios with full-sample portfolios, the study shows that low-centrality strategies combined with the DDPG model perform better in terms of risk control and returns. Furthermore, the comprehensive centrality index evaluation can identify marginally traded stocks with differentiated risk profiles, thereby providing investors in the new energy sector with a scientifically sound and reasonable portfolio strategy.</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

Network-driven DDPG for low-risk energy portfolios

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年7月29日

查看全文
  • Yixuan Zheng,
  • Meihua Wang

摘要

Under the dual challenges of global warming and environmental degradation, the global demand for energy continues to escalate, accompanied by an increasing emphasis on technological innovation. The new energy sector has emerged as a pivotal driver for sustainable development. This study investigates China’s new energy industry through the integration of complex network theory and deep reinforcement learning models. By synergizing complex network analysis with the Deep Deterministic Policy Gradient (DDPG) algorithm, we quantitatively examine risk interdependencies and node significance among listed enterprises using statistical metrics, ultimately aiming to achieve portfolio risk diversification and return enhancement. A multi-layer complex network model is constructed to analyze risk correlations within the new energy sector. A composite evaluation index is developed to identify low-centrality stocks, forming a stock pool characterized by reduced risk interconnectedness. The DDPG algorithm is then applied to optimize portfolio allocation strategies, balancing risk diversification with the dual objectives of maximizing returns and minimizing volatility. In conclusion, this research demonstrates that complex networks can quantitatively describe the degree of interconnection among assets in the new energy sector. By comparing low-centrality portfolios with full-sample portfolios, the study shows that low-centrality strategies combined with the DDPG model perform better in terms of risk control and returns. Furthermore, the comprehensive centrality index evaluation can identify marginally traded stocks with differentiated risk profiles, thereby providing investors in the new energy sector with a scientifically sound and reasonable portfolio strategy.