中文摘要：

针对目前我国城市内河普遍遭到污染的问题,在分析影响内河水质因素的基础上,选取BOD5（5日生化需氧量）、CODcr（化学需氧量）、石油类、挥发酚、NH3-N（氨氮）、总磷等6个主要因素作为评价因子,建立了城市内河水质评价的投影寻踪分析模型。采用模拟退火算法对评价模型进行优化,并将该模型应用于南宁市10条内河水质的评价与排序中。研究表明：用投影寻踪回归分析法进行水质评价,避免了传统评价方法由于主观原因造成的误差,评价结果合理可信、方法简单,具有计算效率高、收敛速度快、受初始条件约束少等优点,为我国城市内河水质的评价提供了新途径。

英文摘要：

In the light of pollution problems of inland rivers in China at present and based on analysis of water quality influential factors,selecting 6 factors including BOD5,CODcr,petroleum,volatile phenol,NH3-N and TP,a water quality evaluation method based on project pursuit analysis is established for water quality evaluation of inland river and it is optimized by Simulated Annealing Algorithms.The model is applied in Nanning City to evaluate and rank 10 inland rivers.The research results show that water quality evaluation by projection pursuit regression analysis avoids the errors caused by subjective reasons in traditional method,and the evaluation result is reliable.The method has the advantages of easy calculation,high calculation effectiveness,fast convergence speed,few constraints of initial condition,and it provides a new way for water quality evaluation of inland rivers.