Miranda, E., C. Ricciardi, J. Suñé, and G. Milano
A SPICE Stochastic Compact Model for Memristive Nanowire Self-Organizing Networks
Based on Jump-Diffusion Langevin Dynamics
IEEE Transactions on Nanotechnology (Nov.2026)
DOI: 10.1109/TNANO.2026.3752555
Abstract: Memristive self-organizing networks based on Agnanowires exhibit stochastic dynamics that can be described by ajump-diffusion Langevin equation driven by a Wiener process anda homogeneous compound Poisson process with Pareto-distributed jump amplitudes. The Langevin equation models theevolution of the network memory state arising from the stochasticformation and rupture of multiple spatially distributed junctions.Continuous fluctuations and discrete switching events arerepresented by the diffusive and jump components of the equation,respectively, with the latter accounting for the observed heavy-tailed statistics and the broad range of switching-eventmagnitudes. We develop and validate a SPICE-compatiblestochastic compact model that captures the resulting networkdynamics using the Method of Elementary Solvers (MES). A term-by-term analysis of the memory-state equation facilitates theinterpretation of the underlying stochastic dynamics. The compactmodel enables self-organizing nanowire networks to beincorporated as reusable SPICE subcircuits for circuit-levelsimulation. The underlying methodology is readily extendable toother systems governed by similar jump-diffusion stochasticdynamics.
Fig: a) Histogram for the data generated by the voltage source V(nor) and N(0,1) distribution.
b) Fast Fourier transform (FFT) for the data generated. The horizontal line indicates the flat region of the spectrum. Cutoff frequency ~1/dt; and spice model script for the jump-diffusion LD process. The memory circuit comprises four current sources (B1-B4). Random number generators fornor (Normal), poi (Poisson), and par (Pareto) are indicated in blue texts.

No comments:
Post a Comment