Optimizing Queue System Performance Through a Comparative Analysis of M/M/1 and M/M/S Models Using Monte Carlo Forecasting (Case Study: Engine Reconditioning at PT. Intidaya Jember)

Abstract

This study aims to optimize the queuing system in the engine reconditioning process at PT. Intidaya Dinamika Sejati by comparing M/M/1 and M/M/S models, supported by Monte Carlo simulation for demand forecasting. The research was motivated by prolonged customer waiting times and uneven machine utilization, which adversely affected productivity and service quality. Data were gathered through direct observations, interviews, and time studies to capture arrival and service parameters. Monte Carlo simulation was employed to model demand variability over six months, with results integrated into discrete-event simulation models built in Arena. The findings revealed that the existing M/M/1 configuration caused significant bottlenecks at Cylgrinding, Honing, and Connecting ROD stations, leading to average waiting times exceeding 150 minutes.

In contrast, implementing an M/M/S model by adding parallel servers at these critical workstations reduced waiting times by over 90%, shortened queue lengths, and balanced machine utilization. This confirms that combining Monte Carlo forecasting with multi-server strategies effectively enhances operational performance and customer satisfaction. Therefore, it is recommended that the company adopt the M/M/S configuration at high-demand stations and conduct regular evaluations to ensure sustained system efficiency.