The limited capacity of electrical energy transmission systems is one of the important factors that must be observed during the operation of electrical networks. If there are consumers in the network with a non-stationary load, then under a certain set of circumstances, unacceptable loads may occur in the system. These overloads can lead to a complete shutdown of the technological process. The work is devoted to the study of the possibility of smoothing peak loads in systems operating several identical machines with non-stationary periodic power. To solve this problem, we propose a mathematical model, with the help of which the original problem is reduced to the problem of minimizing the integral or minimax functionals that estimate the deviation of the total power of machine tools from some stationary value. With the number of machines three or four, the optimization problem can be solved by a simple enumeration over the nodal points of the constructed partition of the set of initial variables. With more machines, this method becomes inefficient. It is also proposed to solve the problem of minimizing two functionals using the statistical method of a swarm of particles. The particle swarm method is slightly inferior to the enumeration method, but it provides much faster calculations. The comparison was made for a task with three and four machines. With more than four machines, the enumeration method becomes inefficient. The particle swarm method has been shown to work effectively with more than four machines. Software has been created to solve these optimization problems.