Abstract
The article considers that the digital transformation of Kazakhstan's agro-industrial complex has created an urgent need for scientifically grounded methods that can optimize marketing strategies under conditions of resource limitations, production seasonality, and heterogeneous consumer behavior. Existing approaches often rely on simplified analytical models that fail to capture the nonlinear and multi-criteria relationships inherent in agricultural marketing. This study addresses that gap by proposing a hybrid decision-support framework that integrates a modified NSGA-III evolutionary algorithm with machine learning techniques to optimize the allocation of marketing budgets across digital and traditional channels. The approach incorporates three key performance criteria—channel efficiency, target audience reach, and marketing cost—while embedding sector-specific constraints related to the AIC of Kazakhstan. A comparative analysis of current digital marketing optimization practices reveals that most methods lack adaptability to sector-specific production cycles, do not sufficiently analyze trade-offs, or fail to integrate machine learning for post-optimization interpretation. The results of the hybrid model show the formation of three distinct clusters of Pareto-optimal strategies, demonstrating the need for differentiated and adaptive budget allocation patterns. The outcomes confirm that the proposed hybrid methodology provides a rigorous scientific basis for designing cost-effective and high-impact marketing strategies within the digital transformation agenda, AIC.