.png){kind=icon}
.png){kind=icon}
Return
Collaborative Offloading for Interacting Users in Cloud-Edge-Terminal Networks
.png){kind=icon}
PRE
AI
DOI：10.1109/JIOT.2025.3625670.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
The rapid growth of Internet of Things devices has led to an increase in computation-intensive and latency-sensitive tasks, making traditional cloud computing insufficient. Cloud-edge-terminal collaboration can enhance offload efficiency by optimizing processing latency, energy consumption, and price cost. However, in real-world networks, computational results often need to be transmitted to multiple users, increasing the offloading complexity. This article proposes a three-tier collaborative offloading architecture for interacting users, considering constraints such as service caching and various types of resources. The optimization problem of computation latency and price cost is modeled as a Markov decision process. To address the problem, we propose a deep reinforcement learning algorithm based on the soft actor-critic framework. Given the discrete-continuous hybrid action space, the algorithm incorporates a dual-head mechanism. After that, transfer learning is incorporated into the training strategy to improve adaptability in dynamic environments. The simulation results demonstrate that the proposed approach outperforms existing performance, convergence, and adaptability methods.
Keywords:
Cloud computing
Servers
Heuristic algorithms
Optimization
Computational modeling
Collaboration
Internet of Things
Adaptation models
Vehicle dynamics
Costs
Cloud-edge-terminal collaboration
interacting users
service caching
soft actor-critic (SAC)
Journal
IF:
8.9
Papers:
1.4W
Citations:
7.8W