COMPARATIVE ANALYSIS OF OPTIMIZATION METHODS BY COST (TIME) COSTS OF FULL FACTOR EXPERIMENT PLANS

Abstract

Relevance. It is proposed to use methods to search for fish schools, monkey searches, jumping frogs for constructing optimal cost (time) experiment plans in the study of technological processes and systems that allow the implementation of an active experiment on them.The purpose of the work is a comparative analysis of these optimization methods for the cost (time) costs of plans for a full factorial experiment.Method. Methods are proposed for constructing the cost-effective (time-consuming) implementation of the experiment planning matrix using fish search, monkey search, jumping frogs algorithms. At the beginning, a number of factors and transition costs are entered for each level of factors. Then, taking into account the entered data, the initial planning matrix of the experiment is formed. The fish search method is based on rearranging the columns of the experiment planning matrix, based on the sum of the values (times) of transitions between the levels for each of the factors. The schools of fish are formed according to the following principle: there are fewer schools of fish where the sum of the values (times) of transition between the levels of factors is greater. Then permutations of fish schools located side by side in the experiment planning matrix are performed. When using the monkey search method, the columns of the experiment planning matrix are trees. Each tree consists of branches along which a monkey moves. There are more tree branches where there is less sum of costs (times) of transitions between levels of factors. The monkey begins its movement upward along each branch of the tree. During this, a search is performed on the branches on which the monkey is located by the minimum value of the sum of the values (times) of transitions between the levels for each of the factors. In the jumping frog method, a successful frog is determined by the least cost of transitions between levels for each of the factors. After this, permutations of frogs are performed. The frog strives for the most successful and, provided it is nearby, it remains in its current location. Then the gain is calculated compared to the initial cost (time) of the experiment.Results. Developed software that implements the proposed methods, which was used to conduct computational experiments to study the properties of these methods in the study of technological processes and systems that allow the implementation of an active experiment on them. Optimum cost plans for the implementation of the experiments were obtained, and the gains in the optimization results compared with the initial cost of the experiment were given. A comparative analysis of optimization methods for the cost (time) costs of plans for a full factorial experiment has been carried out.Conclusions. The experiments have confirmed the performance of the proposed methods and the software implementing them, and also allow us to recommend them for practical use in constructing optimal experiment planning matrices.