ALGORITHMS FOR SOLVING THE PROBLEM OF OPTIMAL PLACEMENT OF DATA IN BILLING OLTP SYSTEMS BASED ON THE IMPLEMENTATION OF THE RANKED APPROACH

Abstract

The article presents the results of the development of approximate and exact algorithms for solving the problem of optimal data placement in billing OLTP systems based on the implementation of the rank approach. This type of problem belongs to the class of integer linear programming (ILP) with Boolean variables. Combinatorial methods and evolutionary algorithms currently occupy a dominant place in the methods of solving such problems. The practical application of these methods is complicated when solving large-scale problems. To eliminate this problem, it is suggested to use the idea of the rank approach. The analysis of approaches to the optimal placement of data in billing OLTP systems, the model of the rank approach, as well as approximate and exact algorithms are given. The object of research is the algorithms of functioning and placement of information resources in the cloud environment, the concept of cloud computing and multi-level information systems. The subject of the study is the principles of data placement stored in distributed databases and circulating in the cloud network, as well as the specifics of subscriber service processes in modern implementations of OLTP systems. The purpose of the research work is to develop approximate and accurate algorithms for optimizing the placement of RDB fragments on network nodes of the cloud structure, which will allow to increase the productivity of the information system due to the rational distribution of data. Conclusions. Strategies for cutting off unpromising paths in sets are proposed, leading to approximate and exact solutions of the problem of integer linear programming with Boolean variables, and efficient exact and approximate algorithms are constructed. It is shown that an important advantage of the developed algorithms based on the rank approach is the fact that an increase in the number of constraints practically does not affect the error of algorithm solutions, while for methods of solving discrete optimization problems based on the ideas of the branch-and-bound method, an increase in the number of constraints to several hundreds actually leads to the impossibility of their practical application.