Research On Quantum Particle Swarm Optimization And Its Application To Codebook Design Of Image Vector Quantization

Quantum intelligent optimization algorithms have better population diversity than their classical counterpart. And the former have better population dispersion, the chromosomes of small population can express multiple search state. Besides, quantum intelligent optimization algorithms have better parallelism, faster convergence speed and much strong global search capability. Particle swarm optimization (PSO), which imitates biological behavior in nature, is a kind of swarm intelligence algorithm. Quantum particle swarm optimization (QPSO), which combines quantum computation with particle swarm optimization, is a classic quantunm intelligent optimization algorithm. This dissertation mainly researches on the optimum performance of the quantum particle swarm optimization and its improved algorithm, and applies the algorithm to the image vector quantization codebook design. The main research works of this dissertation can be summarized as follows:First of all, this paper investigates the characteristics of the basic PSO algorithm and the PSO algorithm with weights and shrinkage factors. And then we investigate the basic principles and algorithm processes of QPSO. The simulation results show that QPSO has better performance than PSO in complex function optimization problems.Secondly, the principle and performance of quantum evolutionary algorithm and the real coded quantum evolutionary algorithm (RCQEA) are investigated in this paper. A novel QPSO (NQPSO) is proposed by introducing the clou of RCQEA to PSO. The proposed NQPSO has been evaluated with respect to efficiency and reliability by benchmark functions. Then we investigate the performance of the proposed NQPSO on solving the 0-1 knapsack problem. The experiments results show the feasibility and efficiency of the NQPSO.Finally, the LBG codebook design algorithm and its improved algorithm for image vector quantization are investigated in this paper. The designing process and flowcharts of vector quantization codebook design algorithm based on standard PSO, QPSO and NQPSO are given. The test images'simulation results show that the codebook designed by NQPSO has better performance than the codebook designed by other algorithms (including QPSO, the LBG and its improved algorithm).