Analysis of the Multi-Objective Control Sequence Optimization Problem in Bivariate Fertilizer Applicators

The bivariate fertilizer applicator (BAF) is a crucial device for precision agriculture, and the optimization of the control sequence optimization (CSO) significantly impacts the performance of variable-rate fertilization (VRF). This study investigates the CSO problem as a multi-objective optimization problem (CSO-MOP) for BFA through the lens of balanced trade-offs among conflicting objectives, including fertilization accuracy, uniformity, and adjustment rapidity. We employed three multi-objective evolutionary algorithms (MOEAs), including NSGA-III, MOEAD-D, and AR-MOEA. To investigate the problem, we solved several instances for different target fertilization rates and selected appropriate evaluation metrics. Finally, we obtained the Pareto set (PS) from each MOEA and conducted a comparative analysis, including the performance of each algorithm in addressing the CSO-MOP, the conflicts between each pair of objectives, and the effects of the optimized control sequences derived from each algorithm on the three objectives.