Abstract
The article considers the problem of multi-criteria optimization of marketing strategies in the agro-industrial complex (AIC) of Kazakhstan using the Non-dominated Sorting Genetic Algorithm (NSGA-II) and machine learning (ML) methods. A combined approach was used, combining optimization on a set of Pareto-optimal solutions, and K-means clustering was used to analyze the results, which made it possible to identify dominant budget allocation strategies between digital, television, radio, print channels and marketing events, such as exhibitions and product presentations. The method was validated both on synthetic data and on real data from Kazakhstan's agro-industrial complex enterprises with an assessment of the quality of solutions through the Hypervolume Indicator, as well as visualization of the 3D Pareto front. The results obtained during the study demonstrate nonlinear relationships between the criteria and identify optimal configurations for various market segments. The practical significance of the study is confirmed by the adaptation of solutions for Kazakhstan's agro-industrial complex enterprises, taking into account the adopted regional restrictions, in particular on the budget for digital marketing and advertising.