试验设计与方法本试验采用五因素(1/2实施)二次回归正交旋转组合设计。以产量为目标函数,在影响目标函数的诸多因素中,选择了亩施纯氮量、亩施P_2O_5量、亩施农肥量、亩播种粒数、播期等五个主要栽培因素为自变量。五个自变量各设五个水平,根据生产实践和试验资料,在编码方法上采取了线性尺度法、以e及3.2为底的对数尺度法的混合式编码方法(见表 Experimental Design and Methods This study used five factors (1/2 implementation) quadratic regression orthogonal rotation combination design. Taking the yield as the objective function, five major cultivation factors such as the amount of nitrogen per mu, the amount of P2O5 per acre, the amount of acreage per acre, the number of seeding per acre and the sowing date were selected as the independent variables among many factors affecting the objective function. The five independent variables are set at five levels. Based on the production practices and experimental data, a linear scaling method is adopted for the coding method, and a hybrid coding method based on the logarithm scale e and 3.2