|
pp. 5795-5814
S&M4652 Research Paper https://doi.org/10.18494/SAM6365 Published: October 20, 2026 Deep Q-network Reinforcement Learning-based Routing Adjustment Mechanism for Carbon Footprint Reduction [PDF] Tin-Yu Wu, Chu-Fu Wang, Wei-Che Chien, Yi-Kai Chen, and Deng-Ying Liu (Received April 2, 2026; Accepted August 31, 2026) Keywords: routing protocol, deep Q network, quality of service, network energy saving and carbon reduction, deep learning
The flourishing development of network applications and humans’ increasing needs on the internet have contributed to a significant increase in not only global data traffic volume (1) but also carbon emissions generated from network services year after year. Therefore, with the public awareness of environmental sustainability, issues related to network energy saving and carbon reduction are receiving considerable attention. At present, the purpose of most studies related to carbon reduction is to reduce energy consumption to reduce carbon emissions. However, reducing carbon emissions is not the only way to save energy. For carbon reduction, the carbon intensity of data transmission across networks can be measured for data transmission strategy adjustment. However, a minimum level of the quality of service (QoS) should be ensured while choosing data transmission paths with low carbon emissions, in case an overemphasis on reducing carbon emissions compromises the QoS. In this study, we organize the connections between QoS and user experience and identify several significant metrics that most affect user experience and their smallest acceptable values. To achieve a routing adjustment mechanism for guaranteeing QoS and carbon reduction, the median life cycle carbon emissions provided by the Intergovernmental Panel on Climate Change (IPCC) is used as the carbon intensity indicator (CII) to find the route with the lowest carbon intensity within the median range.
Corresponding author: Tin-Yu Wu![]() ![]() This work is licensed under a Creative Commons Attribution 4.0 International License. Cite this article Tin-Yu Wu, Chu-Fu Wang, Wei-Che Chien, Yi-Kai Chen, and Deng-Ying Liu, Deep Q-network Reinforcement Learning-based Routing Adjustment Mechanism for Carbon Footprint Reduction, Sens. Mater., Vol. 38, No. 10, 2026, p. 5795-5814. |