.png){kind=icon}
.png){kind=icon}
Return
Reinforcement learning for strategy-proof reliability-aware redundancy task offloading in UAV-enabled Mobile edge computing network
.png){kind=icon}
PRE
AI
DOI：10.1016/j.asoc.2026.115351.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• We consider UAV-enabled mobile edge computing, in which ESs and UAVs jointly provide services to users. Our objective is to maximize the group efficiency, which is the difference between the value and the cost. We consider the hardware failures of ESs and UAVs. The novelty of our work is the development of a new redundancy task offloading model that allows a task to be offloaded to multiple ESs or UAVs. • We consider a dynamic pricing mechanism based on auctions to provide scalable services. We first propose an optimal mechanism based on the Vickrey–Clarke–Groves mechanism. Since the RARTO problem is NP-hard, we then propose a strategyproofness mechanism based on the clock auction. To obtain high-quality solutions, we propose a method that combines reinforcement learning and greedy strategies. • We show that the proposed mechanism achieves individual rationality, consumer sovereignty, strategy-proofness, and group strategy-proofness. That drives the system into an equilibrium in which no user or no group has an incentive to maximize their utility by submitting untruthful bids. We also show that the approximation of the proposed mechanism is , where N is the number of tasks, and is the maximum/minimum margin group efficiency. • The experimental results demonstrate that the proposed mechanism can obtain highquality solutions within a reasonable time.
Keywords:
UAV-enabled mobile edge computing
redundancy task offloading
strategy-proof mechanism
reinforcement learning
group efficiency
Journal
IF:
6.6
Papers:
1.4W
Citations:
4.8W
Organization
No organization information available