We investigate the Λ particle motion and cluster states of using a novel multiple cooling approach guided by the Control Neural Networks method (Ctrl.NN). The numerical results are well reproduced, and the Ctrl.NN method shows rapid convergence concerning the set of the basis states in comparison to traditional generator coordinate method (GCM) calculations. Taking advantage of this merit, we calculate ground state energies and rotational bands of which agree well with experimental observations. By analyzing the density distributions of the main basis, a slight short-range repulsive correlation between Λ particle and clusters in directions perpendicular to the axis of the core is revealed. The Λ particle presents a significant broad distribution as a consequence of this correlation, which we call an “analog -orbit” structure. Furthermore, we prove that this correlation originates from the competition between kinetic energy and Λ-N interaction, as indicated by the energy curves for different configurations of . In the ground states of and , the Λ particle is confined in a cage-like structure of core nuclei formed by two clusters and valence neutrons. These structures lead to the development of more compact Λ particle configurations, and the shrinkage of - relative distance is also well reproduced.

中文翻译：

---

通过神经网络引导微观计算预测[公式省略]的Lambda运动和簇状态


我们研究了使用由控制神经网络方法 (Ctrl.NN) 引导的新型多重冷却方法的 Λ 粒子运动和簇状态。数值结果得到了很好的再现，并且与传统的生成坐标法 (GCM) 计算相比，Ctrl.NN 方法在基本状态集方面表现出快速收敛。利用这一优点，我们计算了基态能量和旋转能带，其与实验观察结果非常吻合。通过分析主基的密度分布，揭示了Λ粒子与团簇在垂直于核轴的方向上存在轻微的短程排斥相关性。由于这种相关性，Λ 粒子呈现出显着的广泛分布，我们称之为“模拟轨道”结构。此外，我们证明这种相关性源于动能和Λ-N相互作用之间的竞争，如不同构型的能量曲线所示。在 和 的基态下，Λ粒子被限制在由两个团簇和价中子形成的核心核的笼状结构中。这些结构导致了更紧凑的Λ粒子构型的发展，并且相对距离的收缩也得到了很好的再现。