In this paper, the problem of identifying the cellular automata (CAs) is considered. We frame and solve this problem in the context of incomplete observations, i.e., prerecorded, incomplete configurations of the system at certain, and unknown time stamps. We consider 1-D, deterministic, two-state CAs only. An identification method based on a genetic algorithm with individuals of variable length is proposed. The experimental results show that the proposed method is highly effective. In addition, connections between the dynamical properties of CAs (Lyapunov exponents and behavioral classes) and the performance of the identification algorithm are established and analyzed.

中文翻译：

---

基于有限时间间隙不完全观测的元胞自动机识别


在本文中，考虑了识别元胞自动机（CA）的问题。我们在不完整观察的背景下构建并解决这个问题，即在某些未知的时间戳上预先记录的、不完整的系统配置。我们仅考虑一维、确定性、二态 CA。提出了一种基于遗传算法的变长个体识别方法。实验结果表明该方法是高效的。此外，还建立并分析了 CA 的动态特性（Lyapunov 指数和行为类别）与识别算法性能之间的联系。