APPLICATION OF THE SHUFFLED FROG-LEAPING ALGORITHM FOR OPTIMIZATION OF PLANS OF FULL FACTOR EXPERIMENT

The purpose of this article is to further develop the methodology for the optimal cost (time) costs of experiment planning, which includes a set of methods for optimizing experiment plans and software and hardware for their implementation. The object of study: the optimization processes for the cost-based plans of multivariate experiments. The subject of research: the cost-optimization method of experimental design plans, based on the use of shuffled frog-leaping method. Experimental research methods are increasingly used to optimize production processes. Planning an experiment allows you to get their mathematical models with minimal cost and time costs. At the same time, a method and a program in the C ++ programming language were developed for constructing optimal or close to optimal plans for a full factorial experiment applying the shuffled frog-leaping algorithm. This allows you to automate the process of solving the problem, reduce the time to develop optimal plans for the experiment, increase the reliability of the results, reduce the time and cost of the experiments. Its effectiveness is shown in comparison with other methods for optimizing multi-factor experimental designs. The efficiency and effectiveness are confirmed by the coincidence or approximation of the optimal plans obtained by this method and the method of complete enumeration. A number of technological objects are presented on which the operability of the developed method and software was tested, namely: fuel consumption in an internal combustion engine, welding of small thickness plates, production of parts by hot stamping, as well as the process of servicing numerically controlled machines. A comparative analysis of the methods for the synthesis of cost-optimal (time) expenditure plans for a full factorial experiment was carried out and the effectiveness of the shuffled frog-leaping method was shown. It is shown that the difficult task of reducing material and time costs when conducting experimental studies can be solved using the proposed method and the software implementing it