Automatic Calibration of a Farm Irrigation Model: A Multi-Modal Optimization Approach

Abstract

In agriculture, plant cultivation requires to take numerous decisions. One of the major problems is the irrigation: a proper irrigation decision has to be made accordingly to the hydric state of the plant and the soil, and the weather prediction. In precision agronomy, this leads to use hydric sensors combined with a numerical model of growth plant model. Such models can not often be tuned by experts. We proposed an automatic parameter calibration of the potato growth model based on data collected in several open fields. As these parameter calibration problem are ill-posed, the associated black-box optimization problem is supposed to be multi-modal. We then compare the performances of two state-of-the-art Evolution Strategies which use different restart mechanisms to automatically tune the set of parameters on different crops, and shows that multi-modal optimization methods may be recommended for such class of optimization problems.

Short term soil moisture forecasts for potato crop farming: A machine learning approach

“…This approach has proved its relevance and robustness, unfortunately mathematical models introduce biases and biological simplifications. Moreover, they are parametric which means they have to be calibrated in order to use them optimally [18,19,20]. The natural or biological phenomena under study involves features with a relationship among them that may be non-linear.…”