![[2602.18797] Carbon-aware decentralized dynamic task offloading in MIMO-MEC networks via multi-agent reinforcement learning](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.18797] Carbon-aware decentralized dynamic task offloading in MIMO-MEC networks via multi-agent reinforcement learning
Summary
This paper presents CADDTO-PPO, a carbon-aware decentralized task offloading framework for MIMO-MEC networks using multi-agent reinforcement learning, addressing challenges in sustainable IoT deployments.
Why It Matters
As the demand for sustainable IoT solutions grows, this research highlights innovative methods to optimize task offloading while minimizing carbon emissions. It addresses critical issues in resource management, making it relevant for developers and researchers focused on green technology and efficient computing.
Key Takeaways
- CADDTO-PPO framework effectively minimizes carbon emissions in MIMO-MEC networks.
- Decentralized execution allows IoT agents to make autonomous decisions based on local observations.
- The proposed system outperforms traditional methods in terms of carbon intensity and packet overflow rates.
Computer Science > Distributed, Parallel, and Cluster Computing arXiv:2602.18797 (cs) [Submitted on 21 Feb 2026] Title:Carbon-aware decentralized dynamic task offloading in MIMO-MEC networks via multi-agent reinforcement learning Authors:Mubshra Zulfiqar, Muhammad Ayzed Mirza, Basit Qureshi View a PDF of the paper titled Carbon-aware decentralized dynamic task offloading in MIMO-MEC networks via multi-agent reinforcement learning, by Mubshra Zulfiqar and Muhammad Ayzed Mirza and Basit Qureshi View PDF HTML (experimental) Abstract:Massive internet of things microservices require integrating renewable energy harvesting into mobile edge computing (MEC) for sustainable eScience infrastructures. Spatiotemporal mismatches between stochastic task arrivals and intermittent green energy along with complex inter-user interference in multi-antenna (MIMO) uplinks complicate real-time resource management. Traditional centralized optimization and off-policy reinforcement learning struggle with scalability and signaling overhead in dense networks. This paper proposes CADDTO-PPO, a carbon-aware decentralized dynamic task offloading framework based on multi-agent proximal policy optimization. The multi-user MIMO-MEC system is modeled as a Decentralized Partially Observable Markov Decision Process (DEC-POMDP) to jointly minimize carbon emissions and buffer latency and energy wastage. A scalable architecture utilizes decentralized execution with parameter sharing (DEPS), which enables autono...
