Global Conference on Mathematics and Applications (GCMA)

This paper presents a comparative study of three sampling techniques for Monte Carlo simulation of a single-server queueing system subject to catastrophes, control times, batch service, and multiple vacations. The proposed simulation framework employs random sampling, Latin Hypercube Sampling, and Refined Descriptive Sampling to estimate the system's main performance measures, particularly for large batch sizes. The methods are evaluated using variance reduction and relative error as comparison criteria. Numerical experiments demonstrate the effectiveness of Refined Descriptive Sampling in improving estimation accuracy and reducing variance compared with Monte Carlo and Latin Hypercube Sampling. The findings highlight the importance of selecting appropriate sampling strategies for the efficient simulation of complex stochastic queueing systems.