A thesis submitted in partial fulfillment of the requirement for the degree of Master of Science

Thesis (M.Sc.) - Asian Institute of Technology, 1995

This study developed a model that could evaluate the performance of the banking computer system in the Bank of Taiwan (BOT). Performance evaluations have been widely studied and several techniques have been developed to suit different aspects. Although many types of models are designed and used, queueing model seems to be more suitable and useful. However, there are still some challenges for evaluators to overcome. Standard queueing model with Poisson process assumptions is widely used in many applications. Unfortunately, it is not suitable to the system in BOT because the number of arrival transactions is a dependent sequence. To depart from independence, the exponential first-order autoregressive sequence (EAR(1)) is employed to generate interarrival sequences in this study. An independent interarrival sequence can be generated by EAR(l) as a special case where first autocorrelation coefficient is set to zero. The simulation results could be the blueprint to evaluate performance of the banking system in BOT when the first autocorrelation coefficient used in EAR(l) is appropriately set. Once some parameters are set, some performance metrics, such as average queueing time, number of servers needed, the number of transactions that have extra-long queueing time and so on, could be obtained.