Organizations: Department of Ports Management and Shipping, National and Kapodistrian University of Athens, 34400 Euboea, Greece · Laskaridis Shipping Co. Ltd, 14562, Kifisia Greece
Smart shipping operations increasingly depend on collaborative AI, yet the underlying data are generated across vessels with uneven connectivity, limited backhaul, and clear commercial sensitivity. In such settings, server-coordinated FL remains a weak systems assumption, depending on a reachable aggregation point and repeated wide-area synchronization, both of which are difficult to guarantee in maritime networks. A serverless gossip approach therefore represents a more natural approach, but existing methods still treat communication mainly as an optimization bottleneck, rather than as a resource that must be managed jointly with carbon cost, reliability, and long-term participation balance. In this context, this paper presents CARGO, a carbon-aware gossip orchestration framework for smart-shipping. CARGO separates learning into a control and a data plane. The data plane performs local optimization with compressed gossip exchange, while the control plane decides, at each round, which vessels should participate, which communication edges should be activated, how aggressively updates should be compressed, and when recovery actions should be triggered. We evaluate CARGO under a predictive-maintenance scenario using operational bulk-carrier engine data and a trace-driven maritime communication protocol that captures client dropout, partial participation, packet loss, and multiple connectivity regimes, derived from mobility-aware vessel interactions. Across the tested stress settings, CARGO consistently remains in the high-accuracy regime while reducing carbon footprint and communication overheads, compared to accuracy-competitive decentralized baselines. Overall, the conducted performance evaluation demonstrates that CARGO is a feasible and practical solution for reliable and resource-conscious maritime AI deployment.
Figures & tables
Method
Central
Peer-to-peer
Payload
Carbon / energy
Unreliable-link
Sparse peer
Smart-ship
coordinator
decentralization
compression
aware scheduling
handling / recovery
communication
target
FedAvg [ 8 ]
✓
–
–
–
–
–
–
D-PSGD [ 12 ]
–
✓
–
–
–
–
–
SGP [ 13 ]
–
✓
–
–
–
–
–
CHOCO-SGD [ 14 ]
–
✓
✓
–
–
–
–
GossipFL [ 15 ]
–
✓
✓
–
–
✓
–
TABLE I: Positioning of representative methods against the main design dimensions targeted by CARGO.
Fig. 1: CARGO control plane. The CARGO Orchestrator combines utility, carbon, and fairness signals to compute per-node scores {gi(t)}i∈Ωt over the available set. These scores drive the Participation Scheduler , while the Topology Scheduler , Carbon-Aware Compressor , and Gossip Mixer determine the activated graph Et , compression policy Ct , and mixing matrix Wt , respectively. The controller output at round t is the decision bundle Dt=(At,Et,Wt,Ct,rt) , where rt denotes the resynchronization flag.
Fig. 2: Round-level scheduling view. Starting from the available-node set Ωt and the candidate edge set Etcand induced by the current topology snapshot, CARGO first selects the active-node set At⊆Ωt and then activates a degree-bounded communication subgraph Et⊆Etcand .
TABLE II: Core engine variables used for ME cylinder EGT forecasting.
Fig. 3: Visualization of maritime gossip topologies across three regimes: (A) Well-connected, (B) mid, and (C) fragmented.
Regime
Δt
dmax
Conn.
Med.
LCR
(min)
(km)
(cnt/rate)
comps
(mean)
Well-connected
30
200
92/144 (0.639)
1
0.883
Mid
30
80
8/144 (0.056)
2
0.692
Fragmented
60
80
5/72 (0.069)
2
0.686
TABLE III: Connectivity statistics for the three topology regimes.
Fig. 5: Packet-loss robustness across connectivity regimes. Performance and resource profiles under increasing packet-loss probability p∈{0,0.05,0.1,0.2} for three topology regimes: fragmented, mid, and well-connected. Top row: final RMSE ( ×10−3 ); middle row: total carbon footprint (g); bottom row: total energy consumption (J). Error bars indicate variability across seeds. The regimes correspond to distinct encounter opportunities derived from the spatiotemporal contact construction (larger communication radius increases connectivity, while coarser temporal binning reduces effective contacts), enabling a controlled evaluation of learning stability and efficiency as network reliability degrades.
Method
RMSE
Carbon (g)
Energy (J)
Comm. (MB)
pd=0.2 , f=0.25
CHOCO
0.0202±0.0076
0.703±0.038
7219±387
146.9±6.7
D-PSGD
0.0094±0.0010
0.715±0.038
7336±387
587.7±26.9
GOSSIP
0.0094±0.0010
0.725±0.039
7448±397
846.0±8.1
SGP
0.0094±0.0010
0.725±0.039
7441±393
587.7±26.9
CARGO
0.0095±0.0006
0.713±0.039
7443±403
470.5±17.9
TABLE IV: Client-stress grid under mid topology.
Fig. 6: Accuracy-efficiency trade-offs under client stress (mid topology). Six panels show R2 versus carbon footprint for availability dropout pd∈{0.2,0.5} and participation fraction f∈{0.25,0.5,1.0} . Circles use the bottom x-axis (carbon, g), while squares use the top x-axis (communication volume, MB), with both markers plotted at the same R2 for a given method and setting. The figure highlights the practical Pareto structure: CHOCO attains lower carbon/communication at the expense of accuracy, whereas CARGO remains in the high-accuracy region while reducing carbon and communication relative to accuracy-competitive decentralized baselines.
Fixed CO 2
Fixed MB
Method
R2
RMSE
R2
RMSE
CHOCO
0.844±0.124
0.0166±0.0063
0.849±0.116
0.0164±0.0060
D-PSGD
0.933±0.026
0.0114±0.0021
0.811±0.031
0.0193±0.0015
GOSSIP
0.934±0.021
0.0113±0.0018
0.935±0.020
0.0113±0.0017
SGP
0.932±0.026
0.0115±0.0020
0.807±0.037
0.0194±0.0018
CARGO
0.941±0.012
0.0108±0.0011
0.934±0.014
0.0114±0.0012
TABLE V: Matched-budget comparison at a shared carbon budget and a shared delivered-communication budget.
AI agents extend large language models from single prompt-response interactions to long-running, goaldirected workflows that issue many model calls, invoke tools, and interact with external environments. These workflows enable tasks such as software repair, data analysis, and experiment management, but their repeated model invocations can incur substantial carbon emissions. This paper characterizes the carbon emissions of OpenClaw agent workloads using WildClawBench, and shows that emissions depend on token consumption, context cache reuse, and the carbon intensity of the grid. Our characterization identifies deadline flexibility as an opportunity for carbon-aware execution: agent tasks can wait for lower-carbon-intensity periods or shift to lower-carbon grids. However, doing so requires handling uncertain execution time for temporal shifting and cached context recomputation during spatial shifting. We present AgentDecarbonizer, a carbon optimizer for AI agents that runs alongside OpenClaw. Given a task prompt and user-specified deadline, AgentDecarbonizer conservatively estimates task duration and selects deadline-feasible execution schedules, while accounting for cache recomputation overhead during spatial shifting. Evaluated on WildClawBench workloads with 60 agent tasks across four grids, AgentDecarbonizer reduces carbon emissions by up to 57.9 % compared with a carbon-agnostic baseline and by up to 37.5 % compared with a baseline that selects the carbon-optimal grid at task start time.
Leyi Yan, Shuangning Li, Sihang Liu
University of Waterloo, Canada · University of Chicago, USA
The growing deployment of large language models (LLMs) makes per-request routing essential for balancing response quality and computational cost across heterogeneous model pools. Current routing methods rarely consider sustainable energy use and CO2 emissions as optimization objectives, despite grid carbon intensity varying by time and region, and models differing significantly in energy consumption. To address this gap, we introduce Green-Aware Routing (GAR), a constrained multi-objective optimization framework that minimizes per-request CO2 emissions subject to explicit accuracy floors and p95-latency service-level objectives (SLOs). GAR employs adaptive constraint optimization through per-dataset floor tuning and incorporates lightweight estimators for correctness, tail latency, and carbon emissions, enabling real-time routing decisions without additional inference passes. We present GAR-PD, a practical online primal-dual routing algorithm for rolling carbon budgets, alongside heuristic variants that achieve high feasibility coverage while limiting accuracy degradation. Comprehensive experiments across standard NLP benchmarks with heterogeneous LLM pools (7B-70B) demonstrate that GAR achieves substantial carbon reductions while maintaining competitive accuracy and p95 latency guarantees, providing a practical, theoretically grounded approach to sustainable LLM inference.
Disha Sheshanarayana, Rajat Subhra Pal, Manjira Sinha +1
Global decarbonisation targets and tightening market pressures demand maritime logistics solutions that are simultaneously efficient, sustainable, and equitable. We introduce EcoFair-CH-MARL, a constrained hierarchical multi-agent reinforcement learning framework that unifies three innovations: (i) a primal-dual budget layer that provably bounds cumulative emissions under stochastic weather and demand; (ii) a fairness-aware reward transformer with dynamically scheduled penalties that enforces max-min cost equity across heterogeneous fleets; and (iii) a two-tier policy architecture that decouples strategic routing from real-time vessel control, enabling linear scaling in agent count. New theoretical results establish O(\sqrt{T}) regret for both constraint violations and fairness loss. Experiments on a high-fidelity maritime digital twin (16 ports, 50 vessels) driven by automatic identification system traces, plus an energy-grid case study, show up to 15% lower emissions, 12% higher through-put, and a 45% fair-cost improvement over state-of-the-art hierarchical and constrained MARL baselines. In addition, EcoFair-CH-MARL achieves stronger equity (lower Gini and higher min-max welfare) than fairness-specific MARL baselines (e.g., SOTO, FEN), and its modular design is compatible with both policy- and value-based learners. EcoFair-CH-MARL therefore advances the feasibility of large-scale, regulation-compliant, and socially responsible multi-agent coordination in safety-critical domains.
Saad Alqithami
Computer Science Department, Al-Baha University, Albaha 65779, Saudi Arabia