Memristive Complex-Valued Neural Network: Chaotic Dynamics, and Its Application to Enhanced Genetic Algorithm

Chaotic systems can enhance stochastic optimization, but their performance is often limited by the dimensionality of the chaos source. This paper introduces a memristive complex-valued Hopfield neural network (MCVHNN) that generates high dimensional complex-valued chaotic dynamics. The model’s complex behaviors, including multi-scroll attractors, are rigorously analyzed and physically validated on an FPGA. Furthermore, the MCVHNN is employed as a chaos generator for a complex valued genetic algorithm (CVGA) in robotic path planning. The CVGA utilizes complex encoding and chaotic geometric operators, demonstrating superior performance in success rate and path quality over conventional methods. This work establishes an integrated framework linking a novel chaotic system to enhanced intelligent optimization.