A knowledge-and-data-driven modeling approach for simulating plant growth: A case study on tomato growth
A knowledge-and-data-driven modeling approach for simulating plant growth: A case study on tomato growth
复制标题
用于模拟植物生长的知识和数据驱动的建模方法：番茄生长案例研究
DOI：
10.1016/j.ecolmodel.2015.06.006
复制
发表时间：
2015-09-24
影响因子：
3.1
通讯作者：
Hu, Bao-Gang
中科院分区：
文献类型：
--
作者：
Fan, Xing-Rong;Kang, Meng-Zhen;Hu, Bao-Gang
提出了一种基于知识和数据驱动的植物生长模拟方法，该方法由两个子模型组成。一个子模型来自所有可用的领域知识，包括来自基于物理或机械模型的所有已知关系；另一个子模型仅由数据构建，而不使用任何领域知识。采用GreenLab模型作为知识驱动子模型，径向基函数网络(RBFN)作为数据驱动子模型。以番茄作物为例进行了植物生长模型的研究。使用了五年来12个温室试验的番茄生长数据集来校准和测试该模型。与现有的知识驱动模型(知识驱动模型，BIC=1215.67)和数据驱动模型(知识驱动模型，BIC=1150.86)相比，提出的知识驱动模型(BIC=1144.36)在预测番茄产量方面表现出一些优势。特别是，KDDM方法能够对不同类型器官的产量提供强有力的预测，包括叶、茎和果实，即使在没有器官观测数据的情况下也是如此。案例研究证实，KDDM方法继承了KDM和DDM方法的优点。文中还讨论了KDDM方法中叠加和复合耦合算子的两种情况。(C)2015爱思唯尔B.V.保留所有权利。
This paper proposes a novel knowledge-and-data-driven modeling (KDDM) approach for simulating plant growth that consists of two submodels. One submodel is derived from all available domain knowledge, including all known relationships from physically based or mechanistic models; the other is constructed solely from data without using any domain knowledge. In this work, a GreenLab model was adopted as the knowledge-driven (KD) submodel and the radial basis function network (RBFN) as the data-driven (DD) submodel. A tomato crop was taken as a case study on plant growth modeling. Tomato growth data sets from twelve greenhouse experiments over five years were used to calibrate and test the model. In comparison with the existing knowledge-driven model (KDM, BIC=1215.67) and data-driven model (DDM, BIC=1150.86), the proposed KDDM approach (BIC=1144.36) presented several benefits in predicting tomato yields. In particular, the KDDM approach is able to provide strong predictions of yields from different types of organs, including leaves, stems, and fruits, even when observational data on the organs are unavailable. The case study confirms that the KDDM approach inherits advantages from both the KDM and DDM approaches. Two cases of superposition and composition coupling operators in the KDDM approach are also discussed. (C) 2015 Elsevier B.V. All rights reserved.