Reinforcement Learning-driven Mechanism Study of Ecological Compensation to Suppress Carbon Lock-in

• Published in: Journal of Applied Science and Engineering Vol. 29; no. 2; pp. 307 - 314
• Main Author: Yongxin Zhou
• Format: Journal Article
• Language: English
• Published: Tamkang University Press 15.06.2026
• Subjects: carbon lock-in; ecological compensation; reinforcement learning
• ISSN: 2708-9967; 2708-9975
• DOI: 10.6180/jase.202602_29(2).0006

Carbon neutrality, as a fundamental goal of global sustainability, is constrained by the carbon lock-in effect, which limits the progress of low-carbon transitions. Ecological compensation mechanisms offer a promising solution by optimizing land use and enhancing carbon sequestration capacity. However, existing approaches often oversimplify ecological processes, restrict decision-making to discrete actions, and lack robustness against environmental uncertainties. To address these limitations, this paper proposes a reinforcement learning framework based on the deep deterministic policy gradient algorithm (RL-MEC-CL), enabling a more precise representation of the dynamic interactions among carbon emissions, carbon sequestration, and land use. Specifically, RL-MEC-CL, in a continuous action space, leverages an actor-critic architecture with experience replay and target networks to optimize compensation strategies adaptively, balancing carbon reduction benefits, sequestration enhancement, and policy costs. Experiment results demonstrate that RL-MEC-CL not only improves the efficiency of ecological compensation strategies but also exhibits strong robustness and adaptability, offering valuable insights for optimizing ecological governance pathways.