Joint multi-user DNN partitioning and task offloading in mobile edge computing

Liao, Zhuofan; Hu, Weibo; Huang, Jiawei; Wang, Jianxin

doi:10.1016/j.adhoc.2023.103156

Cited by 11 publications

(2 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…[23] takes into account the heterogeneity of vehicle computing capabilities and changes in the number of requesters, formulates the optimization problem of dynamic resource allocation, automatically selects the best partition point to minimize the overall delay of all vehicles, and designs a low-complexity chemical reaction optimization algorithm to solve the problem. [24] proposes a partition and offloading scheme for heterogeneous task server systems, establishes a partition and task offloading model for adaptive DNN models, designs the partition point retention (PPR) algorithm, and gives the optimal partition point (OPP) algorithm to find the cost-minimizing best partition point for each ES corresponding to each MD. The above studies fully consider the differences in vehicle task types and divide tasks into two parts to fully utilize the resources of terminal devices and edge servers.…”

Section: Related Workmentioning

confidence: 99%

Digital Twin-Assisted Vehicles Edge Network Computing Offloading Scheme

Zhu,

Tan,

2023

Preprint

View full text Add to dashboard Cite

Traditional vehicle edge computing research has not fully considered the differences between vehicle tasks and edge server computing resources while often ignoring the use of deep reinforcement learning (DRL), which requires a large amount of training data, and DRL is easy to fall into the problem of local optimality. Therefore, this paper proposed a digital twin (DT) -assisted vehicle edge network computation offloading method. For the problem of DRL requiring a large amount of training data, a real-time data acquisition method based on DT is proposed for vehicle tasks and edge nodes. The differentiability of computing resources proposed a computing model based on deep neural network (DNN) partitioning and early exit mechanism, making full use of the computing resources of edge servers and vehicles, reducing the computation delay and transmission delay of tasks, Aiming at the problem that the asynchronous dominant actor-critic (A3C) algorithm in DRL is easy to fall into the local optimal solution, the dynamic baseline and ε-greedy policy are introduced based on A3C and used to select the channel to reduce the transmission delay further. Experimental results show that this scheme outperforms other comparative methods and can effectively reduce the delay of task offloading in vehicular edge computing.

show abstract

Section: Related Workmentioning

confidence: 99%

Digital Twin-Assisted Vehicles Edge Network Computing Offloading Scheme

Zhu,

Tan,

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Opportunistic network forwarding requires nodes to cache data until an encounter with the destination node occurs [17]. Nonetheless, the mobile devices comprising these networks are typically constrained by limited computational power, storage capacity, and energy, frequently resulting in buffer overflows and data loss [18]. Particularly, flooding-based routing algorithms are prone to causing network congestion, node depletion, and delays in data transmission [19].…”

Section: Introductionmentioning

confidence: 99%

Research on Secure Community Opportunity Network Based on Trust Model

Su,

Liang

2024

Future Internet

View full text Add to dashboard Cite

With the innovation of wireless communication technology and the surge of data in mobile networks, traditional routing strategies need to be improved. Given the shortcomings of existing opportunistic routing strategies in transmission performance and security, this paper proposes a community opportunistic routing decision-making method based on the trust model. This algorithm calculates the node’s trust value through the node’s historical forwarding behavior and then calculates the node’s trust value based on the trust model. Thresholds and trust attenuation divide dynamic security communities. For message forwarding, nodes in the security community are prioritized as next-hop relay nodes, thus ensuring that message delivery is always in a safe and reliable environment. On this basis, better relay nodes are further selected for message forwarding based on the node centrality, remaining cache space, and remaining energy, effectively improving the message forwarding efficiency. Through node trust value and community cooperation, safe and efficient data transmission is achieved, thereby improving the transmission performance and security of the network. Through comparison of simulation and opportunistic network routing algorithms, compared with traditional methods, this strategy has the highest transmission success rate of 81% with slightly increased routing overhead, and this algorithm has the lowest average transmission delay.

show abstract