Controller Application of a Multi-Layer Quantum Neural Network with Qubit Neurons

Abstract

This paper investigates a quantum neural network and discusses its application in control systems. A learning-type neural network-based controller that uses a multi-layer quantum neural network having qubit neurons as its information processing unit is proposed. Three learning algorithms; a back-propagation algorithm, a conjugate gradient algorithm and a real-coded genetic algorithm, are investigated to supervise the training of the multi-layer quantum neural network. To evaluate the learning performance and the capability of the quantum neural network-based controller, we conducted computational experiments for controlling a nonlinear discrete-time plant and a nonholonomic system -in this study a two-wheeled robot. The results of computational experiments confirm both the feasibility and the effectiveness of the quantum neural network-based controller and that the real-coded genetic algorithm is suitable for the learning method of the quantum neural network-based controller.

Adaptive-type servo controller utilizing a quantum neural network with qubit neurons

“…To accurately control the plant output utilizing the quantum neural controllers with a higher success rate, an adequate number of qubit neurons are required in the hidden layer of the quantum neural network [21]. The number of qubit neurons in the hidden layer was determined in advance by investigating the success rate of the quantum neural network's learning with respect to the number of qubit neurons.…”

“…The high learning capability of quantum neural networks with qubit neurons has been demonstrated by several basic benchmark tests and applications [16][17][18][19][20]; however, the possibility of applying quantum neural networks to servo-level controller applications has not been adequately investigated. Although servo-level controller applications utilizing the quantum neural network with qubit neurons have been proposed in previous studies [21,22] and the feasibility and effectiveness of the quantum-neural-network-based controller (hereafter called quantum neural controller) have been demonstrated, the learning of the quantum neural network was conducted as an offline process. From the viewpoint of control applications, the online learning of the quantum neural controller, which corresponds to adaptive control, is more practical than the offline learning of the quantum neural controller [23].…”

Quantum neural network-based intelligent controller design for CSTR using modified particle swarm optimization algorithm

“…They showed that the QNN outperforms other types of neural networks in the control of non-linear systems. After presenting preliminaries of neural networks, their performance as controllers has been examined and the results show that the performance of QNNs is better than classical neural networks (Cao and Shang, 2009; Ganjefar et al, 2014, 2015; Takahashi et al, 2012, 2014).…”