DEVELOPMENT OF ALGORITHMS FOR ESTIMATING MEASUREMENT UNCERTAINTY AND SOFTWARE IMPLEMENTING THE CREATION OF ALGORITHMS USING THE MONTE CARLO METHOD

Sharonov Pavel Alexandrovich

In traditional practice, measurement uncertainty is estimated using the linear propagation method (GUM), which assumes the linearity of the model, independence, and normality of the input quantities. However, in real-world problems of system analysis and management, models are often nonlinear, input distributions are asymmetric, and parameters can be correlated. This leads to a distortion of the results, which necessitates the development of algorithms for estimating uncertainty based on the Monte Carlo method and their software implementation. The purpose of the study is to develop and test algorithms for estimating measurement uncertainty based on the Monte Carlo method, which ensure the correct accounting of nonlinear dependencies, correlations and non-standard distributions of input quantities. To achieve this goal, software is being created that implements the proposed algorithms and allows for numerical modeling, analysis of results, and visualization of distributions in system analysis and management tasks.

Keywords:monte Carlo method, standard and expanded uncertainty, input correlation, distributions, software implementation.