Justification of Parameters Modifiable for Genetic Algorithms of Artificial İntelligence for Solving Multi-criteria Optimization Problems

Known analytical methods of mathematical modeling and optimization, when applied to solving the problem of allocating water resources used for irrigation, cannot always provide optimal results. The purpose of the study is to substantiate the parameters of modified GAs, as a direction of combinatorial artificial intelligence, in relation to the problem of optimizing water distribution. Research carried out on the basis of system analysis made it possible to generalize GA modification algorithms (dynamic algorithms, characterized by changes in population volume during the search process; hybrid algorithms based on Lamarck evolution modeling), as well as various classes of adaptive algorithms, characterized by changes in parameters during solution search procedures. The results of the analysis allow us to more reasonably select options for modifying algorithms and/or their parameters for specific tasks, including optimizing the distribution of limited resources for agricultural production carried out in severely arid conditions. Solving the optimization problem using the example of an irrigation water distribution plan under conditions of its scarcity, solved by the GA method due to the significant nonlinearity of the multifactor TF, provides increased efficiency.