A Distributed Algorithm for Task Offloading in Vehicular Networks With Hybrid Fog/Cloud Computing

Liu, Zong-Kai; Dai, Penglin; Xing, Huanlai; Yu, Zhaofei; Zhang, Wei

doi:10.1109/tsmc.2021.3097005

Cited by 50 publications

(22 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In the context of ADMM solving distributed optimization problems examples include [6] and [7], where ADMM was used to solve consensus and sharing problems, respectively. Also, in [8] ADMM is used with particle swarm optimization (PSO) for task offloading in vehicular networks with hybrid fog/cloud computing. Furthermore in [9], ADMM with proximal operator is used to minimize the fixed-point error in a reinforcement learning problem.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Communication-efficient ADMM using Quantization-aware Gaussian Process Regression

Tudela¹,

Wei²,

Nghiem³

2022

Preprint

View full text Add to dashboard Cite

<p> In networks consisting of agents communicating with a central coordinator and working together to solve a global optimization problem in a distributed manner, the agents are often required to solve private proximal minimization subproblems. Such a setting often requires a decomposition method to solve the global distributed problem, resulting in extensive communication overhead. In networks where communication is expensive, it is crucial to reduce the communication overhead of the distributed optimization scheme. Gaussian processes (GPs) are effective at learning the agents’ local proximal operators, thereby reducing the communication between the agents and the coordinator. We propose combining this learning method with adaptive uniform quantization for a hybrid approach that can achieve further communication reduction. In our approach, the GP algorithm is modified to account for the introduced quantization noise statistics due to data quantization. We further improve our approach by introducing an orthogonalization process to the quantizer’s input to address the inherent correlation of the input components. We also use dithering to ensure uncorrelation between the quantizer’s introduced noise and its input. We propose multiple measures to quantify the trade-off between the communication cost reduction and the optimization solution’s accuracy/optimality. Under such metrics, our proposed algorithms can achieve significant communication reduction for distributed optimization with acceptable accuracy, even at low quantization resolutions. This result is demonstrated by simulations of a distributed sharing problem with quadratic cost functions for the agents. </p>

show abstract

Section: Related Workmentioning

confidence: 99%

“…The added non-Gaussian quantization noise invalidates the Gaussian noise assumption of the GP regression expressed in (8). In this case, the regression cannot be a Minimum Mean Square Estimator (MMSE) anymore, so we must compute the conditional mean which requires a more involved computation.…”

Section: Lmmse Regression With Quantizationmentioning

confidence: 99%

Communication-efficient ADMM using Quantization-aware Gaussian Process Regression

Tudela¹,

Wei²,

Nghiem³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Following the expression in (11), the definition of β k i in (7), and that only the terms including β k i contribute to the uncertainty, we get the expression tr(Cov[x k+1 ; ȳk+1 ;…”

Section: B Joint Trace Expressionmentioning

confidence: 99%

“…) is predicted by the GP and this prediction is used by the ADMM algorithm when the coordinator skips a communication round with an agent. This dynamic is expressed in (7) with the variable β k i , where depending on the communication decision, β k i takes the value of ∇f 1/ρ i (z k i ) or its predicted value. In the context of our problem, we replace ∇f 1/ρ i (z k i ) from the expressions in (A.6) and (A.7) with the dynamics defined in (7), giving the ADMM expression…”

Section: Appendix Appendix A: Proof Of Propositionmentioning

confidence: 99%

See 1 more Smart Citation

Optimal Querying for Communication-efficient ADMM using Gaussian Process Regression

Tudela¹,

Nghiem²,

Wei³

2022

Preprint

View full text Add to dashboard Cite

<p>In distributed optimization schemes consisting of a group of agents connected to a central coordinator, the optimization algorithm often involves the agents solving private local proximal minimization sub-problems and exchanging data frequently with the coordinator to solve the global distributed problem. Such schemes usually cause excessive communication costs to the system, leading to the need for communication reduction for scenarios where communication is costly. The integration of Gaussian Processes (GPs) as a learning component to the Alternating Direction Method of Multipliers (ADMM hase been proven successful in learning each agent’s local proximal operators. Consequently, such a learning technique is effective to reduce the communication between the agents and the coordinator. A key element for the integration of GP in the ADMM algorithm is the mechanism upon which the coordinator decides when communication with an agent is required. The decision strategy used in this setting strongly affects the overall communication expenditure reduction and the ADMM’s algorithm performance. For that reason, we construct a general framework presenting a systematic querying mechanism as an optimization problem that balances the ideas ofmaximizinge the communication cost reduction while minimizing the prediction error. Motivated by such a framework, we propose different query strategies using the uncertainty measurement given by GP, to determine if the prediction is reliable enough to skip a communication round. We propose a joint query strategy following a simplification of the general framework that minimizesana L1 norm communication cost constraint by the trace of the joint uncertainty of the ADMM variables which is calculated using all agents’ prediction uncertainty. Additionally, we derive three different decision mechanisms (motivated by a confidence interval analysis) that make their decision by analyzing an uncertainty measurement for each agent individually. We integrate multiple measures to quantify the trade-off between the communication cost reduction and the optimization solution’s accuracy/optimality. The proposed methods can achieve significant communication reduction and good optimization solution accuracy for distributed optimization, as demonstrated by extensive simulations of a distributed sharing problem with quadratic cost functions for the agents. </p>

show abstract

Deep reinforcement learning‐based joint task offloading and resource allocation in multipath transmission vehicular networks

Yin,

Zhang,

Dong

et al. 2024

Trans Emerging Tel Tech

View full text Add to dashboard Cite

Mobile edge computing (MEC) is increasingly being applied to task computation with the continuing increase of computing‐intensive applications in vehicular networks. The MEC can offload the tasks generated by the vehicle to the MEC server to mitigate the computational constraints on the mobile vehicle. However, most task offloading algorithms can not cope with the dynamic distribution of vehicles in vehicular networks. Moreover, most task offloading algorithms do not consider vehicular networks using multipath transmission schemes. Therefore, it remains a challenge to implement efficient joint task offloading and resource allocation (JTORA) algorithms in multipath transmission vehicular networks. In this paper, we introduce a novel cooperative transmission network architecture in multipath transmission vehicular networks. The computational tasks generated by the vehicles can be offloaded to MEC servers deployed near the base station or to cloud servers via multipath transmission. We formulate the task offloading and resource allocation problem as a multi‐objective optimization problem in a multipath transmission vehicular network. To achieve efficient task offloading and resource allocation under cooperative transmission network architecture, we propose a JTORA algorithm for multipath transmission vehicular networks based on deep reinforcement learning (DRL). JTORA converts the optimization problem into a DRL model and obtains a specific optimization policy through model training. Simulation results show that JTORA achieves lower average task delay and lower average task cost than other traditional heuristic algorithms.

show abstract

A Distributed Algorithm for Task Offloading in Vehicular Networks With Hybrid Fog/Cloud Computing

Cited by 50 publications

References 34 publications

Communication-efficient ADMM using Quantization-aware Gaussian Process Regression

Communication-efficient ADMM using Quantization-aware Gaussian Process Regression

Optimal Querying for Communication-efficient ADMM using Gaussian Process Regression

Deep reinforcement learning‐based joint task offloading and resource allocation in multipath transmission vehicular networks

Contact Info

Product

Resources

About