Accepted articles

A.V. Shishulin, A.V. Shishulina
Nizhny Novgorod State Technical University n.a. R.E. Alekseev

Abstract: Complex random number generators that combine rapid pseudorandom sequence generation with hardware entropy sources have numerous practical applications in stochastic process modeling, machine learning, and information security. This paper presents a method for generating pseudorandom sequences based on a physical model of particle diffusion within a nanoscale periodic structure featuring a nonlinear potential and thermal noise, which serve as sources of dynamic chaos. The sequence of thermally activated transitions between potential minima exhibits irregular and chaotic behavior. The sequence generation is achieved through the digitalization of the particle’s stochastic motion along the energy landscape of the nanostructure. Particle diffusion in a nanostructured medium is described by the Langevin equation and is integrating by using the Verlet method. Additionally, a software implementation of the suggested algorithm is provided in the Ruby programming language. The obtained results demonstrate the potential for using a «pseudophysical» approach, which is based on nonlinear potentials of various physical natures, as a viable alternative to a «purely mathematical» methodology in the tasks involving the generation of random number sets.
Keywords: nanostructures, diffusion, nonlinear potential, dynamic chaos, pseudorandom numbers

V.A. Mikhalevsky¹, A.A. Burtsev¹, V.V. Ionin¹, A.A. Nevzorov^1,2, A.V. Kiselev¹, N.N. Eliseev¹, A.A. Lotin^1,3
1 National Research Centre «Kurchatov Institute»
² National University of Science and Technology MISIS
³ Mendeleev University of Chemical Technology

Abstract: This paper presents a modeling study of the electrical resistance in a memristor structure based on the phase-change Ge₂Sb₂Te₅ material. The changes in resistance are driven by some structural transformations in the active region of the memory cell under the influence of electrical control pulses. A novel planar architecture for memristor structures has been demonstrated. Using the simulation data for this architecture, temperature dynamics and phase transitions are analyzed within the framework of the classical Stefan problem. Optimal parameters for the electrical control pulses are identified. The results demonstrate that once the phase transition threshold is reached, the resistance switching time becomes essentially independent of the control parameters. The proposed memristor architecture exhibits advantages in both power efficiency and the capability for multilevel resistance states, positioning it as a key component for next-generation memristive technologies.
Keywords: memristor, chalcogenides, phase change materials, thin films, amorphization