现代回归设计在间作中的应用初探——应用回归设计分析玉米在间作中的密度与方式

APPLICATION OF REGRESSION DESIGN IN INTERPLANTING ——ANALYSIS ON PLANT DENSITY AND PATTERN IN INTERPLANTING OF CORN AND SOYBEAN BY USING REGRESSION DESIGN

本文分析了利用二次回归旋转组合设计,研究玉米大豆间作群体各限制因素对产量等目标性状的影响。通过建立密度、带型配置指数等决策变量与产量、产值等目标函数的数学模型,由计算机进行模拟分析和寻优,得到了如下结论:(1)影响目标性状的主控因子重要程度为玉米密度>带型配置>两者的高互作用。(2)在本试验因子取值范围内,密度在各种带型配置下与产量均呈抛物线关系;而在不同密度下,带型指数对产量的影响为平缓的凹型抛物线关系。利用所建模型,可了解产量产值等目标函数达最大值时相应的决策变量取值,了解全面实施时的所有试验方案的结果,以利优化筛选,为玉米大豆间作的实施和产量、产值的预测预控提供科学依据。

The influence of limiting factors in a population of intercropped corn and soybean upon object characters such as yield was studied with quadratic regression rotational combinatorial design. Mathematical model of object functions such as yield and output value upon strategic variables of plant density and stripe compound index were established. The following conclusions can be made through computerization of simulation and optimization. (1) The sequence in the importance of main controlling factors was plant density of corn stripe compound index and the interaction. (2) Plant density of corn had a paraboal relation with total yield at every stripe compound index and the stripe compound index had a concave parabolic relation with total yield at every plant density in the range of investigated value of the trial factors. With the models established optimization and forecast yield and output value will be possible with the values of strategic variables when objective functions of yield and output value were at maximum and the deduced results of unexperimented trial plant.