Yizhi Xu | Searcharxiv

arXiv2026

Hierarchical Crystal Structure Prediction of Zeolitic Imidazolate Frameworks Using DFT and Machine-Learned Interatomic Potentials

Crystal structure prediction (CSP) is emerging as a powerful method for the computational design of metal-organic frameworks (MOFs). In this article we employ CSP to perform high-throughput exploration of the crystal energy landscape of zinc imidazolate (ZnIm2). As the most polymorphic member of the zeolitic imidazolate framework (ZIF) family, ZnIm2 has at least 24 reported structural and topological forms, and new polymorphs still being regularly discovered. With the aid of custom-trained machine-learned interatomic potentials (MLIPs) we have performed a high-throughput sampling of over 3 million randomly-generated crystal packing arrangements and identified 9609 energy minima characterized by 1484 network topologies, including 855 topologies that have not been reported before. All but one experimentally-reported structures of ZnIm2, falling within the search boundaries, were ultimately matched with the predicted structures, demonstrating the power of the CSP method in sampling experimentally-relevant ZIF structures. Finally, through a combination of topological analysis, density and porosity considerations, we have identified a set of structures representing promising targets for future experimental screening. as well as demonstrated how structures of mechanochemically-synthesized MOFs could be identified via matching experimental powder diffraction patterns with simulated patterns from the predicted structures.

cond-mat.mtrl-sci↗

arXiv2026

CASCADE: A Cascading Architecture for Social Coordination with Controllable Emergence at Low Cost

Creating scalable and believable game societies requires balancing authorial control with computational cost. Existing scripted NPC systems scale efficiently but are often rigid, whereas fully LLM-driven agents can produce richer social behavior at a much higher runtime cost. We present CASCADE, a three-layer architecture for low-cost, controllable social coordination in sandbox-style game worlds. A Macro State Director (Level 1) maintains discrete-time world-state variables and macro-level causal updates, while a modular Coordination Hub decomposes state changes through domain-specific components (e.g., professional and social coordination) and routes the resulting directives to tag-defined groups. Then Tag-Driven NPCs (Level 3) execute responses through behavior trees and local state/utility functions, invoking large language models only for on-demand player-facing interactions. We evaluate CASCADE through multiple micro-scenario prototypes and trace-based analysis, showing how a shared macro event can produce differentiated yet logically constrained NPC behaviors without per-agent prompting in the main simulation loop. CASCADE provides a modular foundation for scalable social simulation and future open-world authoring tools.

cs.HC↗