中文摘要：

在分析光谱颜色再现方法和特点的基础上,结合最小二乘支持向量机小样本、非线性和高维数的特点,提出基于最小二乘支持向量机的颜色光谱重构方法。针对最小二乘支持向量回归机中参数选择的盲目性,结合和声搜索算法具有良好寻优能力和易实现的特点,引入粒子群算法中个体调整参考群体最优值的思想,提出一种全局和声搜索算法,并利用该算法对最小二乘支持向量机中参数进行寻优。实验仿真表明：提出的全局和声搜索算法具有较高的寻优性能,所建立的全局和声搜索优化的最小二乘支持向量机印刷颜色光谱重构方法具有较高的精度。

英文摘要：

Based on the analysis of the characteristics in spectral color reproduce,and combined with the features of small samples,nonlinear and high-dimensional which the least squares support vector machine（LSSVM） has,a model of ink preset in offset printing based on LSSVM was proposed.According to the blindness of parameter choice in LSSVM,and combined with the features of good optimization ability and easy completion of harmony search algorithm,a global harmony search（GHS） algorithm,which brings in the ideal of swarm intelligence,was brought forwarded,and further used to optimize the parameters in LSSVM.The experimental simulation shows that the GHS has a better performance,and the established global harmony search-based least squares support machine（GHS-LSSVM） model has high precision in spectrum rebuild.