A Comparative Study of Rare-Event Simulation Methods for Outage Probability in GSC/MRC Systems under Rician Fading

TL;DR

This paper compares various advanced Monte-Carlo simulation methods for accurately estimating outage probabilities in GSC/MRC systems under Rician fading, introducing a new method and evaluating their efficiency and robustness.

Contribution

It introduces a novel partition importance sampling (PIS) method and compares it with existing techniques for outage probability estimation in Rician fading channels.

Findings

ET, CE, and PIS outperform other methods in efficiency.

CE is the most robust among the studied methods.

Enhanced Monte-Carlo techniques significantly improve outage probability estimation.

Abstract

This paper explores the use of enhanced Monte-Carlo (MC) techniques to evaluate the outage probability of single-input-multiple-output (SIMO) systems under Rician fading, in which the input is combined using generalized selection combining with maximum ratio combining (GSC/MRC). The studied set of methods includes previously established methods: universal importance sampling (UIS) and multilevel splitting (MLS), alongside readapted methods: exponential twisting (ET) and cross-entropy (CE), and a novel method introduced here: partition importance sampling (PIS). Performance is assessed across standard efficiency metrics, revealing that ET, CE, and PIS exhibit the best performance. CE is found to be the most robust method among them.