论文标题
部分可观测时空混沌系统的无模型预测
Dynamical anomalies and structural features of Active Brownian Particles characterised by two repulsive length scales
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。
In this work we study a two-dimensional system composed by Active Brownian Particles (ABPs) interacting via a repulsive potential with two-length-scales, a soft shell and a hard-core. Depending on the ratio between the strength of the soft shell barrier and the activity, we find two regimes: If this ratio is much larger or smaller than 1, the observed behaviour is comparable with ABPs interacting via a single length-scale potential. If this ratio is similar to 1, the two length-scales are relevant for both structure and dynamical properties. On the structural side, when the system exhibits a motility induced phase separation, the dense phase is characterised by new and more complex structures compared with the hexatic phase observed in single length-scale systems. On the dynamical side, as far as we are aware, this is the first representation of an anomalous dynamics in active particles.
