Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Path Assignment in Mesh Networks at the Edge of Wireless Networks

Siddhartha Kumar, Mohammad Hossein Moghaddam, Andreas Wolfgang, Tommy Svensson

TL;DR

This work tackles path assignment in edge mesh networks by introducing a tree-search routing algorithm that accounts for interference across backhaul links and aims to maximize the path SNIR. Path costs are defined as the minimum SNIR along a path, and the optimal path for each user is the one that maximizes this cost, computed via a per-user DFS-tree search over feasible hops up to $h_{\max}$. To maintain scalability, the authors introduce a grouping strategy that partitions users into $G$ groups, reducing the combinatorial search space while trading some optimality. Empirical results show substantial SNIR gains over interference-ignoring and random path selection (3–18 dB and 16–20 dB, respectively) and competitive performance versus a genetic algorithm, with the group-based approach offering favorable complexity characteristics in larger networks. The approach promises improved reliability and throughput for dense edge backhaul deployments, particularly at high carrier frequencies like $60$ GHz.

Abstract

We consider a mesh network at the edge of a wireless network that connects users to the core network via multiple base stations. For this scenario, we present a novel tree-search-based algorithm that strives to identify effective communication path to the core network for each user by maximizing the signal-to-noise-plus-interference ratio (SNIR) along the chosen path. We show that, for three mesh networks of varying sizes, our algorithm selects paths with minimum SNIR values that are 3 dB to 18 dB higher than those obtained through an algorithm that disregards interference within the network, 16 dB to 20 dB higher than those chosen randomly by a random path selection algorithm, and 0.5 dB to 7 dB higher compared to a recently introduced genetic algorithm (GA). Furthermore, we demonstrate that our algorithm has lower computational complexity compared to the GA in networks where its performance is within 2 dB of ours.

Path Assignment in Mesh Networks at the Edge of Wireless Networks

TL;DR

This work tackles path assignment in edge mesh networks by introducing a tree-search routing algorithm that accounts for interference across backhaul links and aims to maximize the path SNIR. Path costs are defined as the minimum SNIR along a path, and the optimal path for each user is the one that maximizes this cost, computed via a per-user DFS-tree search over feasible hops up to . To maintain scalability, the authors introduce a grouping strategy that partitions users into groups, reducing the combinatorial search space while trading some optimality. Empirical results show substantial SNIR gains over interference-ignoring and random path selection (3–18 dB and 16–20 dB, respectively) and competitive performance versus a genetic algorithm, with the group-based approach offering favorable complexity characteristics in larger networks. The approach promises improved reliability and throughput for dense edge backhaul deployments, particularly at high carrier frequencies like GHz.

Abstract

We consider a mesh network at the edge of a wireless network that connects users to the core network via multiple base stations. For this scenario, we present a novel tree-search-based algorithm that strives to identify effective communication path to the core network for each user by maximizing the signal-to-noise-plus-interference ratio (SNIR) along the chosen path. We show that, for three mesh networks of varying sizes, our algorithm selects paths with minimum SNIR values that are 3 dB to 18 dB higher than those obtained through an algorithm that disregards interference within the network, 16 dB to 20 dB higher than those chosen randomly by a random path selection algorithm, and 0.5 dB to 7 dB higher compared to a recently introduced genetic algorithm (GA). Furthermore, we demonstrate that our algorithm has lower computational complexity compared to the GA in networks where its performance is within 2 dB of ours.

Paper Structure

This paper contains 11 sections, 1 theorem, 17 equations, 5 figures, 2 tables, 1 algorithm.

Table of Contents

  1. Introduction
  2. System Model
  3. Link Description
  4. Objective
  5. Tree Search Algorithm
  6. Determining Path with the Lowest Cost in a Tree
  7. Grouping
  8. Complexity Analysis
  9. Results
  10. Conclusion
  11. Proof of Theorem \ref{['thm:time-complexity']}

Key Result

Theorem 1

For a mesh network with $U$ users that are divided into $G$ groups with each group, $j\in[G]$ having $U_j$ users, our group-based algorithm has a time complexity, $\Theta$ that can be upper bounded as where $N_i$ is the number of valid paths for User $i$, and $x$ is the time complexity for calculating the SNIR at a node in the network.

Figures (5)

  • Figure 1: Mesh network with $U=2$ users, $B=7$ BSs and $C=2$ core BSs.
  • Figure 2: A sample interference configuration that occurs at $b_{i_1}$.
  • Figure 3: Initial steps in the construction of the Tree, $T_0$. At each step the node that has a yellow circle in the top right corner represents position of the algorithm. The subsequent node represents the node that the algorithm chooses to explore.
  • Figure 4: Tree, $T_0$ corresponding to the user $u_0$ in the mesh network shown in Fig. \ref{['fig:system_model']}.
  • Figure 5: A mesh network representing 30 BSs, of which 5 are core BS and 15 users that are divided into 6 groups. The gray edges in the network represent the links between the BSs, or a BS and a user. The paths chosen by our grouping based algorithm are represented in red color, while the paths in green color are chosen by Algorithm B.

Theorems & Definitions (2)

  • Example 1
  • Theorem 1