Hardware simulation of semi-Markov and related processes: Part I: a versatile generator

Abstract

After a brief review and classification of the known generators, this part I of 2 parts paper investigates carefully a new programmable generator of general ordinary renewal processes in discrete time. The generator is based on the concept of failure rate and its implementation is completely digital. The effects of time and amplitude quantization are analyzed in detail. Statistical tests, checking distribution and independence, are partly developed and then applied to numerous examples; results are satisfactory. The distributions of the generated random time intervals are not truncated. Thereafter the generator of renewal processes is extended to allow the hardware simulation of general semi-Markov processes in discrete time. (The generation of related processes will be considered in part II, together with applications.)