ЗАСТОСУВАННЯ МЕТОДУ ЗОЗУЛИНОГО ПОШУКУ ДЛЯ ОПТИМІЗАЦІЇ ПЛАНІВ БАГАТОФАКТОРНИХ ЕКСПЕРИМЕНТІВ

Abstract

The methods of experiment planning, optimization, and forecasting are becoming increasingly important in the setting up of studies aimed at studying complex technological processes. The design of experiments involves the incorporation into practice of engineering research of methods to increase the efficiency of work. Unlike the most common one-factor research method, when the effect of each factor is studied separately, there are methods that allow performing experiments in order to study complex processes so that they vary by all factors at once. This helps to increase the efficiency of the experiment, expressed in the fact that the parameters of interest to the experimenter are determined with a significantly smaller error than with traditional research methods. Moreover, with an increase in the number of factors, the accuracy of the experiment increases. Many methods developed in relation to the planning of experiments, making optimal decisions at various stages of research work. But they give a positive result with a small value of the number of factors, because as the number of factors increases, the value of various combinations of permutations increases. Conducting experiments requires certain financial and time costs. Therefore, one of the tasks of optimizing experimental design is to minimize the costs of conducting experimental designs, while obtaining the maximum amount of information about the influence of factors of interest on the process. The purpose of this article is to develop a method and software for optimizing multi-factor experiment designs, which will reduce the time and financial costs of conducting multi-factor experiment designs. An algorithm was developed, which is implemented programmatically, in the C # programming language, to optimize plans for multifactor experiments using the cuckoo search method. The efficiency of the developed algorithm was tested on the optimization of plans for multifactor experiments of technological processes. A comparative analysis of the methods of synthesis of cost-optimal plans of a multi-factor experiment is carried out and the effectiveness of the cuckoo search method is shown.