Transient Memory via Local Pressure Heterogeneity in Non-Equilibrium Systems

Standard physics suggests that chaotic, random systems should forget their past instantly. But is that always true? In this paper, we propose a framework called Dimensional Memory Encoding (DME) to show how out-of-equilibrium systems can actually retain a short-term "memory" of their history.

Through 2D Langevin dynamics simulations of a driven granular gas, we demonstrate that this memory is hidden in local pressure differences.. essentially, the system encodes its past movements into pressure fluctuations. We detected a clear, statistical link between these pressure spots and future particle flow. Finally, we show that this memory isn't free: using a new control parameter ($\Gamma$), we prove that storing information comes with a direct thermodynamic cost. This work provides a fresh, rigorous way to understand how order and memory can emerge from pure chaos.