PSO-PP模型在杨树造纸适宜性综合评价中的应用

Application of PSO-PP Model on Papermaking Suitability Comprehensive Evaluation of Poplar Varieties

选择14种杨树纸浆材品种为研究对象,根据各树种的生长量、纤维形态和化学成分等指标,通过构造投影指标函数进行优化设计,建立基于粒子群算法的投影寻踪模型(PSO-PP)对其进行造纸适宜性评价。该模型根据样本数据类内密集度和类间散开度构建最大目标函数,将多维指标综合转化成一维指标,进而根据相近原则进行分类和排序。结果表明:该模型用于纸浆材造纸适宜性综合评价可行,且与其他方法的分析结果有较好的一致性,为评价纸浆材造纸适宜性提供了一条有效途径。

14 poplar varieties for papermaking wood have been investigated and different variances including volume growth, fiber shape and chemical composition are selected for evaluation. By means of constructing a projectionindex function, a projection pursuit model based on particle swarm optimization （PSO-PP） was applied to study papermaking suitability integrated evaluation of different poplar varieties. Determined by the characteristic of density and variance of the sample datum, PSO-PP can project high dimensional data to low dimensional space through an optimized projecting direction, then the values of the evaluation index were synthesized into a one-dimension projec- tion value. Through studying the main characteristics in 1-D space, the key information can be used to classify and evaluate the samples according to the principle of similarity. The results of case study show that PSO-PP model is feasible in practice in suitability integrated evaluation of papermaking wood, and the analysis results are rightly consistent with the other way, providing an efficient approach in comprehensive evaluation of papermaking suitability.