Quantum Computational Intelligence And Its Application

Quantum Computation (QC) has been one of the advancing fields in modem sceience and technology, which has the properties of parallelism, exponential storage and speed up etc.. It is combined with Computational Intelligence (CI) to develop the Quantum Computational Intelligence (QCI), that openning a new road for CI. Therefore, the research of QCI has the important value both on the theory and practice.In order to solve the key problems of traditional CI, such as the optimum of neuro computational model, the global convergency of particle swarm optimization algorithm and so on, the key technologies of QCI and their applications are researched. The main work and innovation are as follows:(1) Quantum Neuro Network (QNN) technology is researched. Firstly, a new neuron—quantum neuron model is described. Secondly, its logic computational operation and nonlinear property are discussed, that explaining further that the QNN has smaller scale and simpler network topology than neuro network (NN) for signal process. Quantum LM neuro network and its learning algorithm are proposed. The simulation of classifying Iris data base shows that Quantum LM algorithm is superior to traditional LM algorithm, and has notable classified effect.(2) Quantum Evolutionary Computation (QEC) technology is researched. Quantum Particle Swarm Optimization (QPSO) is deduced and discussed in theory. The three typical function optimizations testify its global convergency. The simulation results show that QPSO not only has high convergence speed and good population diversity, but also can conquer the prematurity problem, which is superior to BPSO and GA. A new Quantum Discrete Particle Swarm Optimization (QDPSO) is presented in order to conque the shortcomings of QPSO in discrete problem. Taking advantage of its optimal mechanism, a new CDMA multiuser detection system based on QDPSO is desigened. As the comparisons of simulations demonstrate that the system has good performances of anti-MAI and near-far resistant.(3) LS-SVM technology based on QPSO is researched. To avoid the problem of solving inverse matrix in LS-SVM algorithm, an improved LS-SVM algorithm is presented, the main process is to iteratively solve the linear system of equations with QPSO algorithm. The typical example shows that the idea is feasible and effctive. Furthermore, a gas layer recognition model based on the improved LS-SVM algorithm is desigened, that is to reduce the attributes of samples with rough set method, and then to build the model and recognize with LS-SVM technology based on QPSO. The actual application in certain gas-field indicates that it has the great accuracy and notable applied effect.