PID Parameter Tuning Research Based On Quantum Neural Network

Quantum Neural Network is a new computing model developed by the combination of quantum computation and neural network computation theory. The basic concepts and principles of quantum computing are introduced into the neural network model. The classical neural network model is modified to quantum Neural Network by these concepts and principles. Compared with classical neural network, quantum neural network is greatly enhanced at the convergence performance and rate and put forward new ideas for the further development of neural network.This dissertation gives a comprehensive description of the classical artificial neural network model and the basic concepts and principles of the quantum computing theory. On these basis, the study focuses on the quantum neuron model and Quantum Neural Network. A improved quantum neural neuron model whose output are real variables is proposed based on the principle of the quantum wave function collapse. Then analogized with the classical BP neural network learning algorithm, quantum neural network based on gradient descent learning algorithm is proposed. The main work of the dissertation is as follows:(1) The basic concepts and principles of quantum computing is applied to the neural network and quantum neural network is constructed. The quantum neural neuron is improved based on the quantum superposition of quantum states, conversion and quantum wave function collapse. Input data is processed in quantum neuron using quantum superposition(2) An improved quantum neural network is constructed combined the classical Back Propagation algorithm and the improved quantum neuron. Detailed derivation of the quantum neural network’s learning algorithm is presented based on the gradient descent algorithm.(3) The improved quantum neural network is combined with PID controller and then a quantum neural network based PID controller. Finally, simulation results with Matlab using quantum neural network based PID controller show that the quantum neural network has a strong self-tuning adaptive function.