A crop planning and real-time irrigation method based on site-specific management zones and linear programming

• Two new models to improve the crop planning and real-time irrigation problem.
• These models consider the chemical and physical soil properties and real-time data.
• We use linear programming to efficiently solve the models.
• Experimental results on real and random generated instances validate our method.

We present two mathematical models that consider the physical and chemical site-specific management zones within the parcels. Indeed, the spatial variability of these properties directly impacts the agricultural production planning. The first model is for the crop planning problem. At the beginning of the production cycle, it assesses the chemical and physical management zones to determine the optimal crop pattern for maximizing the farmer’s expected profit. The second model is a real-time irrigation method that takes the solution of the crop planning problem as input. Then, at each irrigation period, it considers the physical management zones and the humidity level of each parcel to determine in real-time the optimal amount of water for each crop irrigation. This is especially important in regions where droughts are frequent. We empirically show that our methodologies are efficient on instances based on real data.