.png){kind=icon}
Return
Matrix-friendly differential evolution algorithm based on memristor
.png){kind=icon}
PRE
AI
DOI：10.1140/epjp/s13360-026-08196-1.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This paper presents a memristor-based circuit design of matrix-friendly differential evolution (MDE) algorithm. By leveraging the unique characteristics of the VTEAM memristor, the variation, crossover, and selection modules of the MDE are constructed. The population input memristor circuit encodes matrix elements using voltage sequences, endowing it with favorable scalability and stability; meanwhile, variations in memristor states do not affect the final output results. The circuit is applied to the optimization of binary non-convex functions. Experimental results demonstrate that the circuit outperforms previous optimization algorithms in terms of iteration count and convergence rate, and can effectively visualize function graphs to assist in the search for global optimal solutions, which underscores the advantages of this method in optimal solution seeking.
Journal
E
IF:
2.9
Papers:
255
Citations:
0