Abstract
The article examines the application of a genetic algorithm to improve the energy efficiency and reliability of pump units operating in parallel. Modeling was carried out in the Matlab&Simulink environment, enabling a detailed analysis of the dynamics of parallel operation of centrifugal pumps. The control inputs were the frequencies of four drives, and the criteria were maintaining the required pressure while minimizing total power and observing hydraulic constraints. A series of computational experiments demonstrated how operator settings and objective-function weights affect convergence speed and solution quality. The results show a stable reduction in energy consumption while preserving operating reliability. Based on the analysis, practical recommendations are proposed for selecting genetic-algorithm parameters for pump-system optimization tasks.