The introduction of active-adaptive devices is one of the key areas in the modernization and reconstruction of the modern electric power industry. This paper proposes and investigates a new effective algorithm for optimal allocation of flexible alternating current transmission systems (FACTS) devices in an electric power system (EPS) of thousands of nodes. The optimal allocation problem is divided into two subtasks: an approximate search for a set of promising installation sites (without analyzing the modes of the power system) and clarifying the found set (based on the analysis of modes). Within the proposed algorithm, the approximate search is a multi-objective optimization problem. A multithreaded implementation of genetic algorithm finds solution for the problem. The result of the algorithm performance is a Pareto front. Each point of the front corresponds to a set of promising installation sites of FACTS. The paper presents the results of the algorithm performance for optimal placement of series FACTS in IEEE 300 bus system. The obtained Pareto sets contain dozens of promising FACTS locations. When number of FACTS increases, the number of points on the Pareto front rises. The paper also considers the implementation of the algorithm for constructing the set of Pareto-optimal solutions based on parallel computing.