and we were interested in analysing daily performance of the appointment system. For analysis of alternate appointment systems to be evaluated, a fully loaded system was considered, i.e., all the 18 time slots were allottedlscheduled in advance to the pa- tients (ie it is assumed that at least 18 requests for appointments are received). The variations over the days of the week were not considered here, and the demand for MRI by a particular class of patient and hence the number of patients of that class to be sched- uled was determined according to the proportion of that patient class in the sample patient population. If a walk-in arrives at a time t such that ti s t ti+1, i e during the ith slot, and (t ti) s 20 helshe is forced to wait until 20 min are past from the slot start time, i.e., ti This ensures that the patient scheduled during this ses- sion does not miss the slot even if he/she is unpunctual (accord- ing to the assumptions, the most unpunctual patients would turn up 20 min late if they show up at all). If (t ti) 20, the walk- in may be served immediately if the server is idle. For each of the appointment systems, 1000 independent replications were carried out so as to keep the half width 95% confidence intervals of the per- formance measures to be less than 5% of the mean performance. The input and output variables for simulation model of OPBEG IBFI