Rational approximation of performance index in stochastic serial production line

Considers the performance evaluation in a stochastic serial production line via rational approximants. Based on the exact or approximate performance indexes of small size systems and their asymptotic behaviors while system size tends to infinity, low order Newton-Pade approximants are used to extrapolate the performance indexes of large size systems which are quite difficult to obtain accurately. This approach is also available to systems with general probability distribution and thus yields a good tool to estimate the performance indexes of those general systems which are hard to deal with by analytical approaches. The simulation examples demonstrate the effectiveness of the approach