论文标题
部分可观测时空混沌系统的无模型预测
On the Expressive Power of the Normal Form for Branching-Time Temporal Logics
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。
With the emerging applications that involve complex distributed systems branching-time specifications are specifically important as they reflect dynamic and non-deterministic nature of such applications. We describe the expressive power of a simple yet powerful branching-time specification framework -- branching-time normal form (BNF), which has been developed as part of clausal resolution for branching-time temporal logics. We show the encoding of Buchi Tree Automata in the language of the normal form, thus representing, syntactically, tree automata in a high-level way. Thus we can treat BNF as a normal form for the latter. These results enable us (1) to translate given problem specifications into the normal form and apply as a verification method a deductive reasoning technique -- the clausal temporal resolution; (2) to apply one of the core components of the resolution method -- the loop searching to extract, syntactically, hidden invariants in a wide range of complex temporal specifications.
