基于混沌万有引力算法对APSIM模型中旱地春小麦产量形成参数的优化

Optimization of Dryland Spring Wheat Yield Formation Parameters inAPSIM Model Based on Chaotic Gravitational Algorithm

为了解决APSIM模型中春小麦产量形成参数本土化率定过程中所面临的耗时长、精度差、效率低等问题,采用混沌万有引力(chaotic gravitational search algorithm, CGSA)算法,基于1971-2014和2018-2021年甘肃省定西市统计年鉴中的产量数据以及2015-2017年定西市安定区凤翔镇安家沟村的大田试验数据、1971-2021年定西市安定区的产量和气象资料,对春小麦产量形成参数进行优化。结果表明,采用CGSA优化参数后,均方根误差(RMSE)、归一化均方根误差(NRMSE)和模型有效性指数(ME)的平均值分别为22.98 kg·hm^(-2)、1.393%和0.995,说明模型在甘肃省定西市春小麦产量的评估中表现出较好的适应性。此外,CGSA具有较好的全局寻优性能和较快的收敛性,为APSIM模型的参数优化提供了一种高效、精准的方法。

Chaotic gravitational search algorithm(CGSA)was introduced to solve the problems of long term,poor precision and low efficiency in the localization of spring wheat yield formation parameters in APSIM model.The algorithm was inspired by the law of gravitation and combined the characteristics of particle swarm optimization.To verify the effectiveness of the algorithm,optimization experiments were carried out based on the yield data from 1971-2014 and 2018-2021 statistical yearbooks,the field test data from 2015-2017 in Anjiagou village,Fengxiang Town,Anding District,Dingxi City,Gansu Province,and the yield and meteorological data from 1971-2021.After using CGSA optimization parameters,the values of RMSE,NRMSE and ME were 22.98 kg·hm^(-2),1.393%and 0.995,respectively.The model showed good adaptability in the evaluation of spring wheat yield in Dingxi City,Gansu Province.In addition,CGSA had better global optimization performance and faster convergence,which provided an efficient and accurate method for parameter optimization of APSIM model.This study provides a new research concept and scientific basis for improving the reliability of model calibration parameters under the conditions of limited field test data.