Abstract
This paper investigates quantum neural networks and discusses its application to controlling systems. Multi-layer quantum neural networks having qubit neurons as its information processing unit are considered and a direct neural network controller using the multi-layer quantum neural networks is proposed. A real-coded genetic algorithm is applied instead of a back-propagation algorithm for the supervised training of the multi-layer quantum neural networks to improve learning performance. To evaluate the capability of the direct quantum neural network controller, computational experiments are conducted for controlling a discrete-time system and a nonholonomic system - in this study a two-wheeled robot. Experimental results confirm the effectiveness of the real-coded genetic algorithm for the training of the quantum neural networks and show both feasibility and robustness of the direct quantum neural control system.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have