Ukhina H. Models and methods to increase the frequency-dependent components efficiency of computer systems.

The work is devoted to solving an urgent scientific and technical problem, increasing the frequency-dependent components (FDC) efficiency, by reducing computational costs during tuning, by analyzing and improving the components model, taking into account the peculiarities of their frequency characteristics and finding their transfer functions coefficients. A study of existing methods for adjusting the frequency-dependent components, assessing the stability and the transient process duration. It is found that these methods implementation requires a large number of operations, and therefore they have high computational costs, which complicates their use in real-time systems. The paper develops new mathematical models of DFC first- and second-order that reflect the frequency characteristics. Based on the obtained mathematical models, a method of restructuring with reduced computational costs due to the conversion of "constants" or an additional value has been developed. The stability assessment and the transition process duration study during the restructuring and developed methods for their assessment. For all developed methods, computational costs are calculated and compared with existing methods. It was found that the estimating stability method with reduced computation is 43% more effective, and estimating the transition process duration is by 28%. The improved tuning method the frequency-dependent components is 27% more efficient for low-pass filter and high-pass filter, 29% - for second-order low-pass and low-pass filter and 13% for SP and BP.