In recent years, there has been extremely active work on a global scale, on a problem that is related to the development of new telecommunication systems, networks and models for them. Optimization using evolutionary algorithms is one of the most successful areas of artificial intelligence, with applications in many areas of engineering education.

The specificity of the rapid change in the development of technologies and equipment makes it difficult to provide a variety of modern equipment in the training of students of technical specialties. This makes it difficult for technical universities to follow the trends of this change. Ways out of this situation are testing in the learning process with experimental and simulation research of test models for optimization in telecommunications.

The authors consider the motivational aspect and the possible training tools and come to the main part of the present report, namely simulation test models for optimization of telecommunication models.

After studying the methods and taking into account the relevance and importance of the problem, giving rise to the need to search for new methods and approaches related to in-depth experimental research, the authors turn their attention to the reasons affecting the quality indicators of telecommunication networks and systems. Simulation test models of communication devices and procedures applying artificial intelligence methods have been created.

The current development is related to the optimization of telecommunication models, by using evolutionary algorithms, applicable in the process of training students of technical specialties.

During the study, the significance of the factors affecting the quality of telecommunication networks were analyzed.

As test reference models in this paper, the authors choose: a test model of a linear system without noise and a test model of a linear system with noise.

The presented test models of telecommunication devices, to which the identification methods are applied, are based on artificial intelligence techniques during the training of engineering students, directly corresponding to their training in communication systems.

The software tools for their implementation have been created and described. The specifics and structure of the software are considered in sequence. The selection criteria and indicators are presented. In order to be used to evaluate the performance of genetic algorithm (GA - Genetic Algorithm) and particle swarm optimization (PSO - Particle Swarm Optimization) algorithms, various linear models of dynamical systems have been prepared.

The following approach (algorithm) was chosen for the entire scientific research, for using the methods and techniques of artificial intelligence, in the evaluation of the components in the telecommunication systems:
- creation of test models of basic telecommunication devices in the Matlab environment;
- testing the created reference models, by simulating and creating input-output;
- creation, in the Matlab environment, of procedures for implementing the selected methods of artificial intelligence;
- performing a comparative analysis and assessment of the quality of evaluation using GA and PSO of the parameters of the test systems.

The described methods are directly related to the training of engineering students who are supposed to start working in a real environment after completing their studies at the university.

Keywords: Telecommunications, artificial intelligence, optimization, engineering learning, test models.