Strategies for Solving the Problem of Representative Sampling of Production Tasks Based on Large Historical Enterprise Data

This study addresses the critical challenge of determining the minimally sufficient volume of initial data necessary for the accurate construction of operation-based technological time norms in the context of processing large-scale historical production datasets. The proposed methodology incorporates advanced techniques for the detection and elimination of erroneous data, as well as systematic data reduction and decomposition achieved through partitioning the data into independent subsets. Given the significant size and complexity of the input datasets, the rationale for leveraging supercomputing resources to solve the resulting high-dimensional optimization problems is thoroughly justified. Moreover, the work explores the implementation of parallel algorithms and high-performance computing technologies, which collectively enhance the computational efficiency and practical applicability of the proposed approach in real-world industrial settings.