AF-TT-RDQN: Ablation Study Results for Maritime Edge Computing Offloading

Experimental results from an ablation study of the action-factorized Rainbow DQN with token Transformer algorithm. The dataset, shared by Shengtian Zhang on figshare under CC-BY-4.0, was last updated on 2026-05-11. It contains results demonstrating the algorithm's performance on a weighted latency-energy objective in an edge computing-enabled Internet of Ships environment.

Use Cases

• Benchmarking new DRL algorithms against AF-TT-RDQN based on the reported weighted latency-energy objective.
• Analyzing the impact of action factorization mechanisms on learning efficiency in high-dimensional decision spaces.
• Validating computation offloading strategies that aim for zero invalid action rates and strict constraint satisfaction.
• Studying the trade-offs between latency and energy consumption in maritime edge computing simulations.

Strengths

• Dataset is openly licensed under CC-BY-4.0.
• The associated study reports a zero rate of invalid actions and satisfaction of all system constraints for the learned strategies.
• The algorithm described outperformed multiple baselines on its primary objective.

Limitations

• The dataset is very small at 5.5 KB, indicating limited scope.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

figshare

Collection Method

Likely contains experimental results from simulations of a proposed deep reinforcement learning algorithm.

Freshness

Last updated 2026-05-11 17:38:00; freshness should be verified.