Modern information technologies require fast performance of algorithms, which can be achieved with the help of parallel com-puting. However, depending on the parameters that determine the characteristics of the subtasks and the mechanisms of their inter-action, the use of parallel computing can lead to both speeding up and slowing down the computations. Testing parallel algorithms in real conditions is resource-consuming, in view of this, the paper proposes a method for optimizing the parameters of parallel computing based on Petri-object modeling and the evolutionary method.Nowadays, there is no unified method for creating a model of parallel computing and, accordingly, there is no tool other than a real program that can be used to optimize the parameters of a parallel program. The lack of simulation tools hinders the development of highly efficient parallel computing. Despite the fact that there are not many tools for simulating parallel computing, there is a movement of research in this direction. The analysis of the existing tools of testing multithreaded programs showed that they are aimed, first of all, at the analysis of the correctness of the execution of computations, but not at the analysis of the efficiency of parallel computing.The research objective is to improve the efficiency of using parallel computing in information technologies by optimizing the parameters of parallel computing based on Petri-object models, which can be used to estimate the execution time of a parallel algo-rithm.The proposed method is based on the simulation of parallel computing by a stochastic Petri net and the application of models in an evolutionary method or a step-by-step optimization algorithm to find parameters that will ensure the efficient performance of parallel computing. As an example of the application of the method, a model of a parallel algorithm for discrete-event system simula-tion was built and investigated. The optimal values of the parameters found using the method quite accurately correspond to those found during experimentation with the parallel algorithm in real conditions.

На швидкодію паралельних алгоритмів сильно впливають параметри, що визначають характеристики підзадач та механізми їх взаємодії, а також обчислювальні ресурси, які використовуються для виконання програми. Тесту-вання паралельних алгоритмів в реальних умовах є ресурсовитратним, з огляду на це у роботі пропонується метод оптимізації параметрів паралельних обчислень на основі Петрі-об’єктного моделювання та еволюційного алгоритму. У якості прикладу застосування методу побудована та досліджена модель паралельного алгоритму імітації дис-кретно-подійної системи.