The proposed universal digital simulators of random processes based on their Markov models are considered as capable of generating sequences of samples of unlimited duration. It is shown that a simple Markov chain allows generating the random numbers with a specified two-dimensional probability distribution of the neighboring values while a doubly connected Markov model makes it possible to get the three-dimensional random numbers. The parameters of the model are determined from either a known probability density or experimental samples of the simulated random process. It is demonstrated that the simulation algorithms do not require complex mathematical transformations and that they can be implemented using a simple element base. To change the properties of the generated random processes one needs to reload the memory device with a pre-formed data array. The block diagrams of the simulators are studied and the probabilistic and correlation characteristics of the generated random processes are determined. It is established that with these simulators a high accuracy of convergence of the probability distributions of the selected model and the histograms of the generated sample sequences is ensured. In the common studies, one can hardly find the results that can surpass by their efficiency the ones that the proposed simulation algorithms demonstrate accounting for their non-problematic hardware implementation (the minimum computational costs) and the simplicity of reconfiguring the Markov model based simulators for generating new random processes. The introduced simulators can be used in the design, development and testing of the multi-purpose electronic equipment, with different meters and the devices for simulating radio paths.
Random-number generator | Matrix of transition probabilities | Markov model | Probability density | Statistical simulation