A key direction in optimizing the operating modes of electric power systems (EPS) is the introduction of series compensation devices (SCDs). The paper addresses the problem of determining the optimal parameters of active-adaptive devices in large-scale EPS. The parameters are determined for a pre-known set of prospective SCD installation sites, represented as a Pareto front and obtained using a multithreaded genetic algorithm. A method for formulating optimization objective functions based on the theory of sensitivity of electrical circuits is presented. The objective functions are formulated using criteria of static and dynamic stability, as well as an analysis of steady-state operating modes of the EPS, which is aimed at comprehensively improving the reliability of the EPS. The practical significance of the approach is demonstrated by the calculation example for a typical IEEE network with 300 nodes. The results are presented in the form of sets of Pareto-optimal solutions for various options for the placement of SCD in the network. Preliminary work requires solving the problem of arranging devices in the EPS, which was previously addressed by the authors.