Distributed Channel Assignment for Ultra-Dense Wireless Networks Using Belief Propagation

Abstract

An efficient channel assignment plays an important role in mitigating co-channel interference in ultra-dense wireless networks. A simple solution is to separate interfering network nodes into orthogonal channels to reduce the interference among them. However, determining the optimal channel assignment is considered to be a non-linear problem, which may also be associated with practical implementation issues such as high computational complexity and control signaling issues. In an effort to cope with these challenging issues, we propose a distributed channel assignment algorithm that efficiently finds the optimal channel configuration by utilizing the concept of belief propagation. Based on a message-passing framework, the proposed distributed channel assignment algorithm maximizes the overall sum rate of the ultra-dense network with a low computational load for each network node. In addition, we design a network protocol and frame format to implement the proposed message-passing framework to real-world wireless networks. The main advantage of the proposed approach is that network nodes autonomously determine the optimal channel assignment and rapidly adapt to dynamic changes of the network. Simulation results confirm that the proposed distributed channel assignment algorithm outperforms conventional algorithms in terms of various network performance aspects, such as the sum rate, scalability, latency, and user mobility.INDEX TERMS Ultra-dense networks, channel assignment, distributed control, message passing, belief propagation.

Distributed Learning Over Markovian Fading Channels for Stable Spectrum Access

“…The same bound can be obtained for (22), and with the same steps, for all terms in (20). The proof for all ∈ M, k̸ ∈G i is similar, and thus Lemma 1 follows.…”

“…Similar bounds can be obtained for the second term in (20). To bound (21) and (22) we use Lezaud's results [57]: Lemma 2 ( [57]).…”

“…The game theoretic aspects of the problem have been investigated from both noncooperative (i.e., each user aims at maximizing an individual utility) [23], [24], [29], [30], [32], [42], and cooperative (i.e., each user aims at maximizing a system-wide global utility) [21], [22], [35], [43], [44] settings. In [19], the distributed allocation problem was investigated under crash and Byzantine failure environments, and the authors in [20] proposed to utilize the concept of belief propagation to efficiently solve the channel assignment algorithm. The notion of access benefit was introduced in [31], where users aim to achieve optimization of their spectrum access by ensuring that the spectrum price will not exceed the access benefits from the spectrum.…”

“…Each of the terms in (20) is the probability of a deviation of the squared difference for two Markov chains' sample means from the squared difference of their expected means by an ϵ. We look at the first term of (20). Using conventional steps from set theory, it can be shown that:…”

AI-Enabled Reliable QoS in Multi-RAT Wireless IoT Networks: Prospects, Challenges, and Future Directions

“…For this purpose, cooperative decision making and federated learning architectures, in which edge node share their learnt information with the central node which regulates the decision of edge nodes, have been proposed in the literature [122], [123], [124], [125]. Various works have tried to address decentralised and distributed network resource management, however, most of these works have focused on routing [125], power allocation [123], channel assignment [126] and scheduling [118], [119]. One of the fundamental requirements to meet QoS requirements at the edge is to understand user's/application's QoS requirements.…”