Simulating Non-Realizable Processes DOI: 10.14209/jcis.2005.26

Consider an output process defined as the sum of M ergodic identically distributed input random processes. Moreover, assume that some output statistics of interest can be found as explicit functions of M with no constraints for M real. Then, extending the investigation of those statistics for non-integer values of M may be desired, mainly if the output process is in effect an idealized approximate model of a realistic random phenomenon. Within this context and for simulation purposes, this work presents a generic, efficient method for generating a sample sequence matching the output statistics for a given noninteger M. To the best of the authors’ knowledge, no solutions to the problem addressed here exist yet.