A Delay Performance Analysis and Wireless Resource Allocation Scheme Based on Martingale Theory

Area of Science:

• Telecommunications Engineering
• Probability Theory
• Network Performance Analysis

Background:

• Statistical delay Quality of Service (QoS) provisioning is critical for Ultra-Reliable Low-Latency Communications (URLLCs).
• Existing frameworks often lack precision in analyzing complex aggregated traffic patterns.
• Efficient wireless resource configuration is essential for meeting stringent latency demands.

Purpose of the Study:

• To develop a precise delay performance analysis framework for URLLCs using martingale theory.
• To design a wireless resource configuration scheme that accounts for statistical delay QoS.
• To establish a method for decoupling QoS requirements into bandwidth demands and optimizing resource allocation.

Main Methods:

• Construction of a martingale for aggregated arrival processes, incorporating heterogeneous bursty and i.i.d. flows.
• Application of martingale stopping time theory to derive the complementary cumulative distribution function of delay.
• Development of a bandwidth estimation algorithm based on theoretical delay violation probability bounds.
• Formulation and solution of a wireless resource allocation problem using Lagrangian convex optimization.

Main Results:

• A tight upper bound for delay violation probability was derived for aggregated traffic.
• The analysis revealed the intricate impacts of entangled heterogeneous flows on system delay.
• A closed-form solution for transmission power was obtained through convex optimization.
• Simulations confirmed the efficacy of the martingale-based analysis and power allocation.

Conclusions:

• The proposed martingale-based framework provides a precise method for analyzing delay performance in URLLCs.
• The developed resource configuration scheme effectively addresses statistical delay QoS requirements.
• The study offers a robust approach for optimizing wireless resource allocation under latency constraints.