论文标题
部分可观测时空混沌系统的无模型预测
HUG model: an interaction point process for Bayesian detection of multiple sources in groundwaters from hydrochemical data
论文作者
论文摘要
储层计算是预测湍流的有力工具，其简单的架构具有处理大型系统的计算效率。然而，其实现通常需要完整的状态向量测量和系统非线性知识。我们使用非线性投影函数将系统测量扩展到高维空间，然后将其输入到储层中以获得预测。我们展示了这种储层计算网络在时空混沌系统上的应用，该系统模拟了湍流的若干特征。我们表明，使用径向基函数作为非线性投影器，即使只有部分观测并且不知道控制方程，也能稳健地捕捉复杂的系统非线性。最后，我们表明，当测量稀疏、不完整且带有噪声，甚至控制方程变得不准确时，我们的网络仍然可以产生相当准确的预测，从而为实际湍流系统的无模型预测铺平了道路。
This paper presents a new interaction point process that integrates geological knowledge for the purpose of automatic sources detection of multiple sources in groundwaters from hydrochemical data. The observations are considered as spatial data, that is a point cloud in a multi-dimensional space of hydrogeochemical parameters. The key hypothesis of this approach is to assume the unknown sources to be the realisation of a point process. The probability density describing the sources distribution is built in order to take into account the multi-dimensional character of the data and specific physical rules. These rules induce a source configuration able to explain the observations. This distribution is completed with prior knowledge regarding the model parameters distributions. The composition of the sources is estimated by the configuration maximising the joint proposed probability density. The method was first calibrated on synthetic data and then tested on real data from hydrothermal systems.
