The Effect of Simulation Parameters on the Selection Approach

AbstractSelection approaches are used to identify the best systemfrom a nite set of alternative systems. If it involves a smallnumber of alternatives, then Ranking and Selection is the rightprocedures to be used in order to select the best system. Nev-ertheless, for a case of a large number of alternatives we needto change our concern from nding the best system to nd-ing a good system with high probability by using the OrdinalOptimization procedure. Almomani and Abdul Rahman [1] hasproposed a new selection approach to select a good system whenthe number of alternatives is very large. In this paper, westudy the eciency of Almomani and Abdul Rahman [1] se-lection approach base on some parameters such as the initialsample size, increment in simulation samples, total budget,and the elapsed (execution) time. In doing so, we apply theirapproach on the M=M=1 queuing systems, in an attempt to de-termine the adequate choices on these parameters in order toget the best performance for the selection approach.