Distributing water optimally is a complex problem that many farmers face yearly, especially in times of drought. In this work, we propose optimization‐based feedback control to improve crop yield and water productivity in agriculture irrigation for a plantation consisting of multiple fields. The interaction between soil, water, crop (sugarcane in this work), and the atmosphere is characterized by an agro‐hydrological model using the crop water productivity modeling software AquaCrop‐OS. To optimally distribute water over the fields, we propose a two‐level optimal control approach. In this approach, the seasonal irrigation planner determines the optimal allocation of water over the fields for the entire growth season to maximize the crop yield, by considering an approximation of the crop productivity function. In addition, the model predictive controller takes care of the daily regulation of the soil moisture, respecting the water distribution decided on by the seasonal planner. To reduce the computational complexity of the daily controller, a mixed‐logic dynamical model is identified based on the AquaCrop‐OS model. This dynamical model incorporates saturation dynamics explicitly to improve model quality. To further improve performance, we create an evapotranspiration model by considering the expected development of the crop over the season using remote‐sensing‐based measurements of the canopy cover. The performance of the two‐level approach is evaluated through a closed‐loop simulation in AquaCrop‐OS of a real sugarcane plantation in Mozambique. Our optimal control approach boosts water productivity by up to 30% compared to local heuristics and can respect water use constraints that arise in times of drought.

中文翻译：

---

大型种植园精准灌溉的优化控制

以最佳方式分配水是许多农民每年面临的一个复杂问题，尤其是在干旱时期。在这项工作中，我们提出了基于优化的反馈控制，以提高由多个田地组成的种植园的农业灌溉中的作物产量和水生产力。土壤、水、作物（本工作中的甘蔗）和大气之间的相互作用通过使用作物水分生产力建模软件 AquaCrop-OS 的农业水文模型来表征。为了在田间最佳地分配水，我们提出了一种两级最佳控制方法。在这种方法中，季节性灌溉规划者通过考虑作物生产力函数的近似值来确定整个生长季节的田间水的最佳分配，以最大限度地提高作物产量。此外，模型预测控制器负责土壤湿度的日常调节，尊重季节性规划者决定的水分分布。为了降低日常控制器的计算复杂度，基于 AquaCrop-OS 模型确定了混合逻辑动力学模型。该动力学模型明确地结合了饱和动力学以提高模型质量。为了进一步提高性能，我们使用基于遥感的冠层覆盖测量值来考虑作物在整个季节的预期发展，从而创建蒸散模型。通过在莫桑比克实际甘蔗种植园的 AquaCrop-OS 中进行闭环模拟来评估两级方法的性能。