Convergence properties of ordinal comparison in the simulation of discrete event dynamic systems (Q1359450)

The paper provides theoretical results explaining the fast convergence of ordinal comparison in the simulation of discrete event dynamic systems. The results include a formulation of an indicator process to characterize the rate of convergence for ordinal comparison and proofs that for several forms of performance measures that are common in simulation, the rate of convergence is exponential. The paper shows that many performance measures of averaging type have asymptotic normal distributions and that ordinal comparison converges montonically in the case of averaging normal variables which is useful for simulation planning.