COMPARATIVE ANALYSIS OF PSEUDORANDOM SEQUENCE GENERATORS AND THEIR EVALUATION

Abstract

The paper investigates the features and performance of pseudorandom sequence generators as essential components of modern information systems used in data processing, simulation, cryptography, and software testing. The study focuses on a comparative analysis of four approaches: the linear congruential generator, the middle-square method, the Mersenne Twister generator, and a generator based on cryptographic hashing. A dedicated software tool was developed to simulate the operation of these generators and to provide their comprehensive evaluation using both statistical and visual analysis techniques. To assess the quality of the generated sequences, the Pearson chi-square test, variance estimation, and autocorrelation analysis were applied. The obtained results are presented in the form of histograms and sequence plots, enabling a clear evaluation of distribution uniformity, randomness, and the presence of hidden patterns in the generated data. The analysis demonstrates that the Mersenne Twister and cryptographic generators exhibit the best statistical properties, including values close to theoretical expectations, minimal autocorrelation, and high uniformity of distribution. It was found that the linear congruent generator can produce acceptable results only with carefully selected parameters, while the middle-square method shows rapid degradation of sequences and poor statistical quality, making it unsuitable for practical applications. The findings confirm the importance of selecting an appropriate generator depending on the application domain and required statistical or cryptographic properties. The proposed comprehensive evaluation approach can be effectively applied in both educational and practical contexts for analyzing and comparing pseudorandom number generators.