A revisit of two-stage selection procedures

Two-stage indifference-zone selection procedures have been widely studied and applied to determine the sample sizes for selecting a good system among k alternative systems. Dudewicz and Dalal (1975) present two selection procedures: P E and P R. The derivation of the P E procedure is based on the weighted sample means while P R is based on the overall sample means. In this paper, we provide the rationale that P (CS∣P E) ⩽ P (CS∣P R). We then carry out simulation experiments to compare View the MathML sourceP(CS∣PE,S12(n0),S22(n0),…,Sk2(n0)) and View the MathML sourceP(CS∣PR,S12(n0),S22(n0),…,Sk2(n0)) when k > 2. Furthermore, we propose a new approach to estimate the critical constant for Rinott’s (1978) procedure.