Effective control of simulation runs by a new evaluation algorithm for correlated random sequences

A new algorithm is presented which allows the analysis of correlated random sequences in order to gain the stationary d.f. Fn( z). This algorithm being based on recent statistical investigations of Markov chains can be used to control objectively the required number of trials n of a computer simulation run by a formula which depends on the desired minimum value Fmin of Fn( z), on the prescribed upper limit c:lmaz of relative error and also on the measured mean value of the correlation coefficient p. As shown by selected examples the algorithm analyzes any correlated z-sequence of discrete and/ or continuous type and may replace the conventional batch means evaluation method and other methods.