中文摘要：

针对综合考虑产量、品质、水分利用效率等指标的灌水模式优选问题，本文提出了融入加速遗传算法的膜下滴灌灌水模式投影寻踪评价模型。该模型通过高维数据向低维空间的转换，把样本的多个评价指标转化成一个综合指标，实现了对膜下滴灌多种灌水模式的评价。最后应用该模型于新疆的大田试验中，对9种灌水模式进行了评价和优选，结果表明，对整体考虑产量、品质、水分利用效率的最优灌水模式的选择与大田实验方案选择相符，能客观、真实地反映各评价指标的贡献率和方向性。

英文摘要：

The projection pursuit evaluation model is applied to establish the assessment model for optimizing the irrigation scheme of under mulch drip irrigation for cotton comprehensively considering the indexes of yield, quality and water use efficiency. The real coding based accelerating genetic algorithm is introduced to accelerate the calculation process. This method is used to assess the various irrigation schemes for the field experiment of cotton in a station in the Xinjiang Autonomous. It shows that the optimized scheme base on the assessment accords with that obtained from statistic data.