MEMRISTOR MODELING AND SIMULATION USING MATLAB SIMULINK

Memristors—non-volatile elements capable of retaining memory without power—have the potential to revolutionize future memory systems and neuromorphic computing. Despite their theoretical appeal and initial experimental success by HP Labs in 2008, several critical challenges have prevented their widespread adoption. Fabrication inconsistencies, long-term reliability, and difficulties in integrating memristors with existing technologies remain significant hurdles. Moreover, existing SPICE-based models, though valuable, often lack the flexibility to simulate essential parameters such as the normalized doped region width (w/D), non-linear dopant drift, and dynamic behaviors under varying frequencies and voltages. These limitations in literature motivate the need for improved, accessible, and more comprehensive modeling tools.

To address this gap, this research focuses on the modeling and simulation of memristor devices using MATLAB Simulink, which offers a more intuitive environment for visualizing and analyzing device behavior. We examine and simulate key memristor models—including the HP, TiO₂, Joglekar, and VTEAM models—to capture non-linear and time-dependent characteristics fundamental to memristor functionality. Specific attention is given to modeling the hysteresis loops, current-voltage relationships, and the impact of the w/D ratio, which is often overlooked in prior studies.

By bridging the gap between theoretical models and practical simulation tools, this study contributes to a deeper understanding of memristor behavior and modeling strategies. These insights may guide the development of more reliable and efficient memristor-based technologies, enabling their integration into real-world applications such as resistive memory systems and neuromorphic circuits.