A SEVEN-STAGE METHODOLOGY AND TASK BANK FOR TEACHING MATHEMATICAL MODELING OF AGROTECHNICAL PROBLEMS WITH AI-ASSISTED TOOLS

This article presents a seven-stage methodology and an accompanying task bank, comprising 48 tasks (45 individual problems and three integration projects) organized into four thematic sections, for teaching mathematical modeling of agrotechnical engineering problems with AI-assisted tools. A five-criterion selection system - pedagogical effectiveness, professional context relevance, usability, financial-technical feasibility, and integration capability was applied to justify five core AI tools: Maple, Wolfram Alpha, GeoGebra AI, Python/Google Colab, and MATLAB/Simulink. The methodology's effectiveness was verified through a multi-site quasi-experimental study conducted across three higher education institutions (N=436) using Welch's t-test, ANCOVA, and MANOVA (p<0.001, η²=0.72), while replication stability across sites was statistically confirmed (ICC=0.03).