Universal Stasis Theory (UST): A Unified Equilibrium-Based Framework for Galactic Dynamics, Cosmology, and Modified Structure Formation

Pattern Persistence Theory (PPT) proposes a unified computational framework in which system identity is defined as the persistence of structural patterns under the joint influence of entropy, memory, and resilience dynamics.
The theory introduces a measurable quantity called Persistence Capacity (Π), which integrates structural robustness (Γ), temporal memory (M), and entropy (Sₙ) into a single dynamical metric governing system stability.
We implement and evaluate this framework using deep neural networks trained on MNIST, where Π is observed to correlate with model accuracy, generalization behavior, and collapse (overfitting) dynamics.
This work bridges theoretical system identity modeling with empirical deep learning behavior, providing a testable and computationally grounded approach to studying stability in complex adaptive systems.