FPLinQ: A cooperative spectrum sharing strategy for device-to-device communications

Shen, Kaiming; Yu, Wei

doi:10.1109/isit.2017.8006944

Cited by 69 publications

(107 citation statements)

References 6 publications

Supporting

Mentioning

107

Contrasting

Order By: Relevance

“…4. Outputs from the convolutional stage are then input into a DNN that learns from the scheduling results of the state-of-the-art FPLinQ algorithm [59] as in [56] or in an unsupervised manner as in [57] to be discussed later. The simulation results suggest that DNNs can effectively learn the network interference topology and perform scheduling to near optimum without the help of accurate CSI.…”

Section: A Supervised Learning Approach For Optimizationmentioning

confidence: 99%

Deep-Learning-Based Wireless Resource Allocation With Application to Vehicular Networks

et al. 2020

View full text Add to dashboard Cite

It has been a long-held belief that judicious resource allocation is critical to mitigating interference, improving network efficiency, and ultimately optimizing wireless communication performance. The traditional wisdom is to explicitly formulate resource allocation as an optimization problem and then exploit mathematical programming to solve the problem to a certain level of optimality. Nonetheless, as wireless networks become increasingly diverse and complex, e.g., in the highmobility vehicular networks, the current design methodologies face significant challenges and thus call for rethinking of the traditional design philosophy. Meanwhile, deep learning, with many success stories in various disciplines, represents a promising alternative due to its remarkable power to leverage data for problem solving. In this paper, we discuss the key motivations and roadblocks of using deep learning for wireless resource allocation with application to vehicular networks. We review major recent studies that mobilize the deep learning philosophy in wireless resource allocation and achieve impressive results. We first discuss deep learning assisted optimization for resource allocation. We then highlight the deep reinforcement learning approach to address resource allocation problems that are difficult to handle in the traditional optimization framework. We also identify some research directions that deserve further investigation.

show abstract

Section: A Supervised Learning Approach For Optimizationmentioning

confidence: 99%

Deep-Learning-Based Wireless Resource Allocation With Application to Vehicular Networks

et al. 2020

View full text Add to dashboard Cite

show abstract

“…α, β, and γ are the Lagrangian multipliers for the constraints (P1c), (P4b) and (P1e), respectively. In (30), the constraints (P1c), (P4b) and (P1e) are written in vector form as y − F z = 0, v − G y = 0, and H y − h = 0, respectively. The positive parameter ρ controls the weight on the quadratic penalty terms, which also corresponds to the step size of the dual descent update in the ADMM based solution to be introduced.…”

Section: A Distributedmentioning

confidence: 99%

“…The positive parameter ρ controls the weight on the quadratic penalty terms, which also corresponds to the step size of the dual descent update in the ADMM based solution to be introduced. We only consider the dual variables of the equality constraints (P1c), (P4b), and (P1e) in (30). The rest of the constraints are omitted here for simplicity, which will be considered when solving each subproblem.…”

Section: A Distributedmentioning

confidence: 99%

“…In each iteration, there are three steps to update the primal and dual variables. First, update y t+1 based on v t , z t , h t α t , β t and γ t calculated from the previous iteration to minimize (30). Then, update v t+1 , z t+1 and h t+1 based on y t+1 , α t , β t and γ t .…”

Section: Y Z and Hmentioning

confidence: 99%

“…In related work [30], the problem is to select an active set of links (equivalently, a pattern) on a particular time slot (equivalently, a frequency band) in a peer-to-peer network to maximize a weighted sum rate. An iterative algorithm based on fractional programming determines a single pattern for each time slot.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Large-Scale Spectrum Allocation for Cellular Networks via Sparse Optimization

Zhuang¹,

Guo²,

Wei³

et al. 2018

IEEE Trans. Signal Process.

View full text Add to dashboard Cite

This paper studies joint spectrum allocation and user association in large heterogeneous cellular networks. The objective is to maximize some network utility function based on given traffic statistics collected over a slow timescale, conceived to be seconds to minutes. A key challenge is scalability: interference across cells creates dependencies across the entire network, making the optimization problem computationally challenging as the size of the network becomes large. A suboptimal solution is presented, which performs well in networks consisting of one hundred access points (APs) serving several hundred user devices. This is achieved by optimizing over local overlapping neighborhoods, defined by interference conditions, and by exploiting the sparsity of a globally optimal solution. Specifically, with a total of k user devices in the entire network, it suffices to divide the spectrum into k segments, where each segment is mapped to a particular set, or pattern, of active APs within each local neighborhood. The problem is then to find a mapping of segments to patterns, and to optimize the widths of the segments. A convex relaxation is proposed for this, which relies on a re-weighted 1 approximation of an 0 constraint, and is used to enforce the mapping of a unique pattern to each spectrum segment. A distributed implementation based on alternating direction method of multipliers (ADMM) is also proposed. Numerical comparisons with benchmark schemes show that the proposed method achieves a substantial increase in achievable throughput and/or reduction in the average packet delay.

show abstract