Research Of The Swarm Intelligence Algorithm On Vector Quantization And TSP

The swarm intelligence algorithm and the theory of vector quantization(VQ) data compression are summarized systemically in the paper. In particular, the current research status and development history of VQ codebook design and codword search are also introduced. The classical algorithm of VQ codebook design—LBG algorithm and traditional codeword search algorithms are reviewed. The swarm intelligence algorithm on combination problems such as TSP are discussed. To overcome the limitation of classical algorithm of VQ codebook design, an improved algorithm that combines the swarm intelligence algorithm with the LBG algorithm is proposed. The simulation results show the validity of the improved algorithm in image compression coding. The traditional codeword search algorithm—full algorithm(FS) has a serious disadvantage, whose computation quantity will rise sharply with the increase of the codebook size and the codeword dimension. This paper proposes an effective algorithm which implements in Hadamard domain, and the Chebyshev distortion measure and PDS algorithm are adopted. The experiment shows the improved algorithm can search codeword faster and effective.The swarm intelligence algorithm includes ant colony optimization(ACO) and particle swarm optimization(PSO). ACO and PSO have high robustness and parallel characters. Because LBG algorithm can find the local optimal solution fastly and PSO has the ability of searching the global optimal solution, a new optimal strategy is proposed in the paper. A better codebook can be designed using the new optimal strategy. To study the characters of the swarm intelligence algorithm, comparison with generation algorithm(GA) on TSP shows the advantage of the swarm intelligence. Finally, the paper works are summarized, and some farther research suggestions regarding VQ and the swarm intelligence algorithm are provided.