论文标题
部分可观测时空混沌系统的无模型预测
Weak supersymmetric $su(N|1)$ quantum systems
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。
We present several examples of supersymmetric quantum mechanical systems with weak superalgebra $su(N|1)$. One of them is the weak $su(N|1)$ oscillator. It has a singlet ground state, $N +1$ degenerate states at the first excited level, etc. Starting from the level $k = N+1$, the system has complete supersymmetric multiplets at each level involving $2^N$ degenerate states. Due to the fact that the supermultiplets are not complete for $k \leq N$, the Witten index represents a nontrivial function of $β$. This system can be deformed with keeping the algebra intact. The index is invariant under such deformation. The deformed system is not exactly solved, but the invariance of the index implies that the energies of the states at the first $N$ levels of the spectrum are not shifted, and we are dealing with a quasi-exactly solvable system. Another system represents a weak generalisation of the superconformal mechanics with $N$ complex supercharges. Also in this case, starting from a certain energy, the spectrum involves only complete supersymmetric $2^N$-plets. (There also exist normalizable states with lower energies, but they do not have normalizable superpartners. To keep supersymmetry, we have to eliminate these states.)