Application rules for destructive ability of genetic algorithm operators in the problem of structural and parametric synthesis of business process simulation models

This study proposes the application rules for destructive ability of genetic algorithm operators in the problem of structural and parametric synthesis of business process simulation models. The aim of the research is to confirm the hypothesis that it is possible to influence the performance of a genetic algorithm by changing the operating parameters of its operators, which allows increasing the convergence of this evolutionary procedure and helps the intelligent algorithm overcome “bottlenecks”. The “bottleneck” of a genetic algorithm is understood as attenuation of the algorithm, finding the population at local extrema of the fitness function, etc. Based on this hypothesis, it is proposed to use an add-on in the form of an artificial neural network to intervene in the process of finding solutions as a control model. It is planned to simulate this process using the mathematical apparatus of Petri nets theory. When implementing such an approach to solving the problem, it is necessary to consider the influence of the destructive ability of operators on the behavior of the population and determine the order of actions that need to be performed to control the evolutionary search for solutions in the problem of structural and parametric synthesis of dynamic business process simulation models. The paper discusses examples of population states of a genetic algorithm as well as the results of applying the proposed rules for making adjustments to operator activities. The main operators that significantly influence the state of the population are considered: the selection operator, the crossing operator, and the mutation operator; the influence of the reduction operator was not regarded in this study.