Carbon-aware decentralized dynamic task offloading in MIMO-MEC networks via multi-agent reinforcement learning

Mubshra Zulfiqar; Muhammad Ayzed Mirza; Basit Qureshi

cs.DCFeb 21, 2026

Carbon-aware decentralized dynamic task offloading in MIMO-MEC networks via multi-agent reinforcement learning

Authors: Mubshra Zulfiqar, Muhammad Ayzed Mirza, Basit Qureshi

Organizations: School of Computer Science and Artificial Intelligence, Wuhan University of Technology, Wuhan, 430000, Hubei, China · School of Computer Science and Information Engineering, Qilu Institute of Technology, Jinan, 250200, Shandong, China · Department of Computer Science, Prince Sultan University, Riyadh, 12435, Riyadh Province, Saudi Arabia

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 autonomous IoT agents to make fine-grained power control and offloading decisions based solely on local observations. Additionally, a carbon-first reward structure adaptively prioritizes green time slots for data transmission to decouple system throughput from grid-dependent carbon footprints. Finally, experimental results demonstrate CADDTO-PPO outperforms deep deterministic policy gradient (DDPG) and lyapunov-based baselines. The framework achieves the lowest carbon intensity and maintains near-zero packet overflow rates under extreme traffic loads. Architectural profiling validates the framework to demonstrate a constant $O(1)$ inference complexity and theoretical lightweight feasibility for future generation sustainable IoT deployments.

Figures & tables

Articles	Approach		MIMO MEC/EH	Objectives (Min.)			Arch Profiling.	Solution Type	Key Contributions & Constraints
Articles	Cent.	Decent.	MIMO MEC/EH	Carbon	Latency	Energy	Arch Profiling.
This work	–	$\checkmark$	$\checkmark$	$\checkmark$	$\checkmark$	$\checkmark$	$\checkmark$	CADDTO-PPO	Sustainable MIMO-MEC; Constraints: Renewable energy, MIMO interference
[ 21 ]	–	$\checkmark$	$\times$	$\times$	$\checkmark$	$\times$	$\times$	Fed. Learning	Secure partial offloading in FL; Constraints: Data privacy, security
[ 29 ]	$\checkmark$	–	$\times$	$\times$	$\times$	$\checkmark$	$\times$	MINLP/Greedy	Elastic time-slot allocation; Constraints: Device battery budget
[ 20 ]	$\checkmark$	–	$\times$	$\checkmark$	$\times$	$\times$	$\times$	Lyapunov	Carbon-aware ML offloading; Constraints: Carbon emission budget
[ 18 ]	$\checkmark$	–	$\times$	$\times$	$\checkmark$	$\times$	$\times$	DQN	Scheduling dependent task graphs; Constraints: Task dependencies

Table 1: Comparison of carbon-aware task offloading and resource management in MEC

Figure 1: Framework of the carbon-aware MIMO-MEC system

Notation	Description
$U$	Set of IoT users
$T$	Index set of time slots
$\delta$	Length of one time slot
$c_{u}(t)$	Processor computing speed for local execution during time slot $t$ of user $u$
$N_{u}(C_{u})$	CPU cycles per task bit
$p_{l,u}(t)$	Local power of user $u$ for local execution at slot $t$

Table 2: Notations and descriptions

Figure 2: System architecture of the multi-user uplink MIMO-MEC system. The BS monitors SINR and provides feedback to decentralized agents for autonomous power control.

Figure 3: Multi-agent PPO architecture with centralized training decentralized execution using shared parameters

Figure 4: Schematic representation of the local observation and decentralized decision-making process of user agent at slot $t$

Network and Communication Settings
Parameter	Value
Cell Radius ( $R$ )	$100$ m
Time Slot Duration ( $\delta$ )	$0.01$ s (10 ms)
System Bandwidth ( $W$ )	$1$ MHz
Noise Power ( $\sigma^{2}$ )	$1\times 10^{-12}$ W
Path Loss Exponent ( $\alpha$ )	$2.0$

Table 3: System simulation and algorithm parameters

Figure 5: Training convergence of CADDTO-PPO and Centralized-PPO.

Panel of Figure 6 — Figure 6: Performance evaluation under varying mean data arrival rates ( $\lambda_{u}\in[2,10]$ bits).

Panel of Figure 7 — Figure 7: Sustainability analysis under varying mean energy harvest rates ( $\lambda_{E}\in[1,5]$ Joules).

Panel of Figure 8 — Figure 8: Scalability analysis across varying user densities ( $U\in[5,50]$ ).

Figure 9: Eco-efficiency trade-off analysis: System throughput vs. carbon intensity.

Metric	CADDTO-PPO	Lyapunov-DPP
Network architecture	MLP [128, 128]	Logic Search
Total parameters	17.28 K	N/A
MAC operations	17.02 K	N/A
FLOP operations	34.05 K	N/A
Storage size (ONNX)	8.29 KB	N/A
Runtime RAM usage	67.51 KB	Low

Table 4: Comparative hardware profiling and deployment metrics

Computation Offloading

See all 15 papers

Apr 28, 2026

QAROO: AI-Driven Online Task Offloading for Energy-Efficient and Sustainable MEC Networks

With the rapid advancement of artificial intelligence (AI) and intelligent science, intelligent edge computing has been widely adopted. However, the limitations of traditional methods, such as poor adaptability and the slow convergence of heuristic algorithms, are becoming increasingly evident. To enable sustainable and resource-efficient edge applications, this paper proposes an online task offloading framework for wireless powered mobile edge computing (MEC) networks, called Quantum Attention-based Reinforcement learning for Online Offloading (QAROO). The system employs a binary offloading strategy with the aim of co-optimizing computing and energy resources in dynamic channel environments. In response to the issues of poor adaptability in traditional approaches and the slow convergence of heuristic algorithms, the framework integrates quantum neural networks and attention mechanisms, introducing three key improvements: using recurrent neural networks to enhance temporal modeling capability, proposing an uncertainty-guided quantization method to improve exploration efficiency, and incorporating attention mechanisms into quantum networks to strengthen feature representation. Experiments demonstrate that the proposed method outperforms comparative schemes in terms of normalized computation speed and processing time, offering an efficient and stable solution for online task offloading in large-scale Internet of Things (IoT) dynamic environments.

Yongtao Yao, Yao Yang, Haorui Shi +3

Edge Computing Wireless Resource Allocation Attention Mechanisms Quantum RL

Ask PDF

Apr 8, 2026

Multi-Turn Reasoning LLMs for Task Offloading in Mobile Edge Computing

Emerging computation-intensive applications impose stringent latency requirements on resource-constrained mobile devices. Mobile Edge Computing (MEC) addresses this challenge through task offloading. However, designing effective policies remains difficult due to dynamic task arrivals, time-varying channels, and the spatio-temporal coupling of server queues. Conventional heuristics lack adaptability, while Deep Reinforcement Learning (DRL) suffers from limited generalization and architectural rigidity, requiring retraining when network topology changes. Although Large Language Models (LLMs) offer semantic reasoning capabilities, standard Supervised Fine-Tuning (SFT) yields myopic policies that greedily minimize immediate latency without accounting for long-term system evolution. To address these limitations, we propose COMLLM, a generative framework that enables foresighted decision-making in MEC systems. COMLLM integrates Group Relative Policy Optimization (GRPO) with a Look-Ahead Collaborative Simulation (LACS) mechanism, which performs multi-step Monte Carlo rollouts while jointly modeling server queue dynamics. By incorporating these rollouts into the reward design, the framework captures the long-term impact of current decisions on future system states. Experimental results demonstrate that COMLLM achieves near-optimal latency and improved load-balancing fairness. Notably, it exhibits zero-shot topological scalability, allowing a model trained on small-scale networks to generalize to larger, unseen topologies without retraining, outperforming SFT, DRL, and heuristic baselines.

Ning Yang, Chuangxin Cheng, Haijun Zhang

Institute of Automation, Chinese Academy of Sciences, Beijing, 100190, China · School of Mechano-Electronic Engineering, Xidian University, Xi’an, 710071, China · School of Computer and Communication Engineering, University of Science and Technology Beijing, Beijing, 100083, China

Edge Computing LLM Reasoning RL for LLM Agents

Ask PDF

Jul 3, 2026

Hierarchical Multi-Agent Reinforcement Learning for Carbon-Aware AI Data Centers in Power Distribution Systems

Eco-friendly energy management for artificial intelligence data centers (AIDCs) is crucial because of the significant increase in energy consumption-induced carbon emissions from AIDCs resulting from the rapid expansion of AI applications. This paper proposes a hierarchical carbon-aware multi-agent reinforcement learning (CA-MARL) framework for robust and efficient operations of AIDCs under uncertainties while ensuring low-carbon operation of power distribution systems. The framework comprises a workload manager (WM) agent and multiple local AIDC agents trained using a multi-agent transformer method, corresponding to a global AIDC aggregator and a local AIDC operator, respectively. Leveraging AIDC operation data along with nodal carbon intensity (NCI) calculated from the carbon emission flow-integrated distribution system operator problem, the WM agent spatially allocates AI training and inference jobs among all AIDCs. Based on the jobs allocated from the WM agent and NCI information, each AIDC agent schedules economical and eco-friendly operations of the AIDC by performing the following tasks: i) temporal shifting of training jobs, ii) spatial allocation of training graphics processing unit (GPU) blocks and inference GPUs within the AIDC, and iii) control of the supply air temperature of the cooling system. The effectiveness of the proposed framework was assessed using an IEEE 33-node power distribution system.

Hyunsoo Lee, Panggah Prabawa, Dae-Hyun Choi +1

LG Electronics, 17709, Pyeongtaek, Gyeonggi-do, Republic of Korea · School of Electrical and Electronics Engineering, Chung-Ang University, Seoul 06974, Republic of Korea · Department of Electrical and Computer Engineering, Korea University, Seoul 02841, Republic of Korea

Energy-Efficient ML Data Center Energy Management Carbon-Aware Computing Multi-Agent Reinforcement Learning Power Systems

Ask PDF

Abstract

Computation Offloading

Explore similar work

QAROO: AI-Driven Online Task Offloading for Energy-Efficient and Sustainable MEC Networks

Multi-Turn Reasoning LLMs for Task Offloading in Mobile Edge Computing

Hierarchical Multi-Agent Reinforcement Learning for Carbon-Aware AI Data Centers in Power Distribution Systems