Hierarchical Model and Decision Optimization Algorithm for Distributed Data Storage and Processing

The task of optimizing distributed data storage and processing is difficult to solve in a limited time. In this regard, a hierarchical approach has been applied to solve it, which provides for the presentation of a generalized problem in the form of a set of hierarchically ordered subtasks, for each of which locally optimal solutions are determined at the appropriate hierarchy level. To optimize solutions for distributed data storage and processing, a process model has been formed, presented in the form of a set of hierarchically ordered components, a mathematical model of a hierarchical game, which is a way to optimize solutions at hierarchy levels. In order to determine effective solutions at hierarchy levels, an algorithm for local optimization of solutions based on genetic algorithms has been developed. The construction of data processing schedules assigned to computing devices is implemented using the proposed heuristic procedure. The application of the developed models of the distributed data storage and processing process, hierarchical game models and algorithms for optimizing solutions made it possible to significantly increase the dimension of the problem, take into account the parameters characterizing data transmission channels when optimizing solutions at hierarchy levels, and minimize the amount of unused resources.

hierarchical game, restrictions on the volume of distributed storage devices, optimization of solutions for distributed storage and distributed data processing, genetic algorithms