Development of bandwidth allocation scheme in wireless communication networks using the Shapley game theory

Abstract

ABSTRACT

Bandwidth allocation is essential for efficient wireless network performance and user satisfaction. Traditional methods like Minimum-Maximum, Proportional allocation, and Weighted Fair Queuing often fail under dynamic conditions, causing unfairness and inefficiency. This study introduces a new bandwidth allocation approach using non-competitive bankruptcy game theory. Users (nodes) request cache space and bandwidth based on their needs, treated as claimants in a system with limited bandwidth. When demand exceeds supply, the Shapley value allocates bandwidth fairly based on individual contributions. Network slicing was used to create virtual networks, each dedicated to specific services and allocated bandwidth using the bankruptcy model, guided by Quality of Service (QoS) parameters like delay, throughput, and reliability. Cache memory was allocated from the kernel to reduce latency and enhance performance. Simulated in MATLAB R2023a, the model’s performance was measured using QoS and bandwidth fairness and compared to Minimum-Maximum and Proportional methods. With bandwidth demands ranging from 500 to 6000 Mbps and 10 to 20 slices, the Shapley method showed 99.7% fairness, 97% QoS efficiency, and only 0.000315 variance. In contrast, Minimum-Maximum averaged 77% fairness with the highest variance of 0.075479. The Shapley method proved most effective, practical, and adaptable for modern bandwidth-limited wireless networks.