A Modified Statistical Process Control Approach for Warm-up Selection in Discrete-Event Simulation
HLA-based simulation systems are multivariate discrete-event simulation systems, and sometimes with small samples in essential, for which exists initial transient issue. Simulation models begin in fixed initial values, which are always different from the steady state values. A variety of selection methods have been developed these years, but no one method could be recommended above others. Among these methods, the familiar statistical process control method is a popular way to deal with the warm-up selection, with the deficiency for multivariate and small samples. Based on SPC, individuals and moving range, multivariate exponentially weighted moving average chart for Individuals and Hotelling T2 method, a modified SPC method is proposed. The case of military domain simulation scenario and some empirical experiments, illustrate the validity of the modified SPC method.
P. Shi et al., "A Modified Statistical Process Control Approach for Warm-up Selection in Discrete-Event Simulation", Advanced Materials Research, Vol. 187, pp. 428-433, 2011