ทันเอไอ

Researchers Propose 'Memory Engine': A Physics Model Where Systems Self-Organize Through Internal 'Memory'

A new study published in Physical Review E introduces a model where particles self-organize into orderly movement patterns by leaving traces on the surface they traverse, acting as the system's 'memory.'

A research team has published a pivotal paper in the journal Physical Review E titled "Memory engine: Self-organized coherence from internal feedback," introducing the 'Memory Engine.' This physics model explains how seemingly chaotic systems can self-organize from within, without external forces or controllers.

At the heart of this model is the concept of "memory" derived from the interaction between particles and their environment. Researchers simulated Brownian particles moving across a special viscoelastic substrate. This surface acts as more than just an empty space; it records the particle's path as a "memory field."

The mechanism is a simple yet powerful internal feedback loop: as a particle moves, it leaves a trace or alters the surface properties. These recorded traces subsequently influence or constrain the particle's future movement. This process operates under the Coupled Memory Graph Process (CMGP) framework, where the system is non-Markovian, meaning it retains its history.

Computer simulations reveal a striking phenomenon: the system shifts from unstructured diffusion to orderly, cyclic movement patterns called "burst-trap cycles," where the particle alternates between rapid movement and temporary entrapment. The key factor governing this transition is the stiffness of the surface serving as the memory unit.