Population-Input Memristor Circuit Based on Matrix-Friendly Genetic Algorithm

This paper presents the design and implementation of a novel matrix-friendly genetic algorithm (MGA) based population input memristor circuit. Selection, crossover and mutation operations can be efficiently implemented in MGA by utilizing the unique properties of memristors. The proposed circuit performs well in solving binary nonconvex function optimization problems. Through extensive simulations and comparisons with classical optimization algorithms, our method improves about 4.24% on average in exploring the optimal fitness value, which reflects the advantages of the method in finding the optimal solution. In addition, the existence of its function to visualize the image of a one-dimensional function is also found during the experimental process. The proposed circuit has good scalability and flexibility, demonstrating its potential for a wide range of applications in complex optimization tasks.