Mohammad H. Almomani, Adam Baharum,
Rosmanjawati Abdul Rahman
School of Mathematical Sciences
Universiti Sains Malaysia
Penang, 11800, MALAYSIA

Abstract. The article studies the effect of the initial simulation sample size on the performance of Three-Stage selection approach. The aim of the study was to see the effects of the initial simulation sample size on the selection approach in selecting the best simulated designs when the number of alternatives system is large. In Three-Stage selection approach, the ordinal optimization technique is used in the first stage to select a subset that overlap with the set of the best designs with high probability, and then subset selection procedure is used to get a smaller subset that contains the best among the subset that is selected by the first stage. Finally, the indifference-zone procedure is used to select the best design among the survivors in the second stage. In fulfilling the aim of the study, the Three-Stage approach was implemented on M/M/1 queuing system under some parameter settings, with a different choice of the initial simulation sample sizes. The results show that the choice of the initial sample size does affect the performance of Three-Stage selection approach.

Received: June 17, 2011

AMS Subject Classification: 62F07, 90C06, 49K30, 49K35

Key Words and Phrases: ranking and selection, ordinal optimization, subset selection, indifference-zone, simulation optimization, initial simulation sample size

Download paper from here.