Multi-Criteria Optimization in the Design of High-Load Systems

Abstract

The current stage of development of science and technology is characterized by a significant complication of the tasks. The development of the economy to produce a situation where the development, implementation and operation of complex technical and socio-technical systems have to be in conditions of fierce competition. This is necessary to reduce the time of development and implementation of new technologies, especially in high-load systems. Highly loaded systems are, by and large, the same websites, only with a very large audience, and as a consequence with a large load, which requires an optimized server part of the site. A qualitative characteristic for a highly loaded system is the bandwidth of this system, it describes the amount of work that must be able to perform the system per unit time. The development of highly loaded systems is time consuming and poorly formalized. Since the production of high-load systems is one of the most dynamically developing areas in the field of information technology, which is demonstrated by a significant annual increase in volumes. In the practical task of decision-making, there is often a situation where you cannot limit yourself to considering a single criterion for choosing a decision. An attempt at mathematical formalization of such problems has led to the creation of the Theory of Multicriteria Optimization, which is used in the development of methods, intended for support of decision makers, in the presence of several criteria. The steps for construction of algorithm of multicriteria optimization in designing of highly loaded systems, carries out search of solutions of a maximum in a multicriteria problem are resulted. The algorithm was based on the descent method in Simplex problems. Simplex method - an algorithm for solving the optimization problem of linear programming by searching the vertices of a convex polyhedron in multidimensional space. Multicriteria optimization is based on finding solutions in problems with a large number of options. Now the type of tasks is very resource-intensive and is calculated using computers. Keywords: high-load data systems, high-load computing system.