Wait time analysis to optimize number of opened counters / Noraini Noordin, Nurul Elfieqah Rostam and Nur Ariena Farhana Noor Hamizan

• Main Authors: Noordin, Noraini, Rostam, Nurul Elfieqah, Noor Hamizan, Nur Ariena Farhana
• Format: Article
• Language: English
• Published: Universiti Teknologi MARA, Cawangan Pulau Pinang 2020
• Subjects: Analysis; Computer simulation
• Online Access: http://ir.uitm.edu.my/id/eprint/40551/1/40551.pdf; http://ir.uitm.edu.my/id/eprint/40551/

The Morning Shift (MS) flow of passengers on Thursday is similar to any other day in the week at an urban train station. However, passenger congestion occurring in the Evening Shift (ES) affects the system behaviour. The system has been showing this characteristic over the years. However, only one counter is opened. This counter also sells different types of tickets. Thus, the system is not cost-effective. The study aims to determine the optimal number of counters that should be opened on Thursday. In order to solve this problem, the study has applied Poisson Queuing Simulation (PQ) to the MS and ES data. Findings indicate that running one or two counters in MS maintains the mean of wait time in-queue and in-system at less than one minute, while the mean of in-queue and in-system passengers is also at one person. Extra cost has to be incurred to hire another teller; thus, one counter is optimal. For ES, the service rate was only slightly higher than the arrival rate. Based on the mean number of in-queue and in-system passengers, there was no way that one counter can efficiently service the queue that was formed. A simulation was done to determine if there was a need to set up a two-counter or a three-counter system. Although a two-counter system will be idle 47% of the time, it was not cost-effective for the management to two extra tellers for a three-counter system. The management must take future corrective measures. Based on the findings, one counter is optimal for MS, but cost-effectiveness tests must confirm that two counters are optimal for ES. Besides, the management should also identify actions that can be taken during the 47% idle system time in ES