草业学报 ›› 2023, Vol. 32 ›› Issue (12): 77-89.DOI: 10.11686/cyxb2023064
收稿日期:
2023-03-02
修回日期:
2023-05-04
出版日期:
2023-12-20
发布日期:
2023-10-18
通讯作者:
郝勇
作者简介:
Corresponding author. E-mail: 518004@yangtzeu.edu.cn基金资助:
Jun-lin LIU1(), Yong HAO1(), Chun-yan LIU2, Yan SHAO1, Lang DING1
Received:
2023-03-02
Revised:
2023-05-04
Online:
2023-12-20
Published:
2023-10-18
Contact:
Yong HAO
摘要:
为了解决BP神经网络在对含根土抗剪强度的预测中存在的预测精度低,计算收敛速度较慢,容易陷入局部极值等问题,本研究通过直剪试验、团聚试验、根系分形分析试验等探究了不同因素对含根土抗剪强度的影响,并对各因素进行相关性分析,从中选取了大团聚体含量(R0.25)、平均重量直径(MWD)、几何平均直径(GMD)、土壤分形维数(Dd)、根表面积、平均直径6个影响含根土抗剪强度的因素作为模型输入层节点,含根土的抗剪强度作为输出层节点。参考FangfaGorman 理论公式、Kolmogorov理论公式以及一种经验公式分别计算,并对结果进行讨论,确定了本研究中神经网络的最佳隐含层节点数量为13。建立6∶13∶1的BP神经网络模型,并引入了烟花算法(FWA)对BP神经网络进行优化。结果显示,BP神经网络、粒子群算法(PSO)优化的BP神经网络、FWA-BP神经网络的预测值与期望值的最大相对误差分别为11.12%、9.06%、7.44%,平均相对误差分别为4.60%、3.24%、1.96%,相较于BP神经网络和PSO-BP神经网络,FWA-BP神经网络预测误差值有明显降低;对比引入的统计参数,均方根误差(RMSE)、平均绝对误差(MAE)、平均绝对百分比误差(MAPE),FWA-BP神经网络对应的数值分别为0.779353、0.625993、5.6679%,均小于作为对比的普通BP神经网络和PSO-BP神经网络,显示出FWA-BP神经网络在实际应用中的优越性。
刘俊麟, 郝勇, 刘春艳, 邵严, 丁琅. 基于烟花算法的BP神经网络预测含根土抗剪强度[J]. 草业学报, 2023, 32(12): 77-89.
Jun-lin LIU, Yong HAO, Chun-yan LIU, Yan SHAO, Lang DING. Prediction of shear strength of root-containing soil by a back-propagation neural network optimized with the fireworks algorithm[J]. Acta Prataculturae Sinica, 2023, 32(12): 77-89.
组 | 平均重量直径 MWD (mm) | 大团聚体含量 R0.25 (%) | 几何平均直径 GMD (mm) | 土壤分形维数 Dd | 根表面积 | 平均直径 | 黏聚力 Cohesion (c, kPa) |
---|---|---|---|---|---|---|---|
1 | 1.252 | 0.856 | 1.082 | 2.342 | 23.527 | 0.728 | 13.6 |
2 | 1.224 | 0.849 | 1.074 | 2.329 | 22.706 | 0.729 | 13.3 |
3 | 1.364 | 0.898 | 1.130 | 2.310 | 25.505 | 0.498 | 14.6 |
4 | 1.315 | 0.879 | 1.110 | 2.317 | 24.213 | 0.462 | 14.5 |
5 | 1.105 | 0.831 | 0.981 | 2.357 | 11.468 | 0.396 | 12.9 |
6 | 1.070 | 0.811 | 0.982 | 2.345 | 11.561 | 0.308 | 11.8 |
7 | 0.905 | 0.741 | 0.964 | 2.433 | 3.645 | 0.285 | 10.2 |
8 | 0.910 | 0.736 | 0.972 | 2.423 | 3.321 | 0.204 | 10.6 |
9 | 1.231 | 0.867 | 1.042 | 2.365 | 22.244 | 0.676 | 12.9 |
10 | 1.285 | 0.858 | 1.041 | 2.345 | 21.702 | 0.762 | 13.8 |
? | ? | ? | ? | ? | ? | ? | ? |
46 | 0.929 | 0.723 | 0.975 | 1.805 | 2.221 | 0.261 | 11.2 |
47 | 1.233 | 0.813 | 1.014 | 1.479 | 23.455 | 0.794 | 13.8 |
48 | 1.321 | 0.874 | 1.189 | 1.769 | 24.899 | 0.459 | 14.2 |
49 | 1.058 | 0.821 | 0.981 | 1.729 | 10.845 | 0.359 | 12.5 |
50 | 0.919 | 0.718 | 0.970 | 1.798 | 2.525 | 0.276 | 10.9 |
表1 训练样本
Table 1 Training samples
组 | 平均重量直径 MWD (mm) | 大团聚体含量 R0.25 (%) | 几何平均直径 GMD (mm) | 土壤分形维数 Dd | 根表面积 | 平均直径 | 黏聚力 Cohesion (c, kPa) |
---|---|---|---|---|---|---|---|
1 | 1.252 | 0.856 | 1.082 | 2.342 | 23.527 | 0.728 | 13.6 |
2 | 1.224 | 0.849 | 1.074 | 2.329 | 22.706 | 0.729 | 13.3 |
3 | 1.364 | 0.898 | 1.130 | 2.310 | 25.505 | 0.498 | 14.6 |
4 | 1.315 | 0.879 | 1.110 | 2.317 | 24.213 | 0.462 | 14.5 |
5 | 1.105 | 0.831 | 0.981 | 2.357 | 11.468 | 0.396 | 12.9 |
6 | 1.070 | 0.811 | 0.982 | 2.345 | 11.561 | 0.308 | 11.8 |
7 | 0.905 | 0.741 | 0.964 | 2.433 | 3.645 | 0.285 | 10.2 |
8 | 0.910 | 0.736 | 0.972 | 2.423 | 3.321 | 0.204 | 10.6 |
9 | 1.231 | 0.867 | 1.042 | 2.365 | 22.244 | 0.676 | 12.9 |
10 | 1.285 | 0.858 | 1.041 | 2.345 | 21.702 | 0.762 | 13.8 |
? | ? | ? | ? | ? | ? | ? | ? |
46 | 0.929 | 0.723 | 0.975 | 1.805 | 2.221 | 0.261 | 11.2 |
47 | 1.233 | 0.813 | 1.014 | 1.479 | 23.455 | 0.794 | 13.8 |
48 | 1.321 | 0.874 | 1.189 | 1.769 | 24.899 | 0.459 | 14.2 |
49 | 1.058 | 0.821 | 0.981 | 1.729 | 10.845 | 0.359 | 12.5 |
50 | 0.919 | 0.718 | 0.970 | 1.798 | 2.525 | 0.276 | 10.9 |
组 | 平均重量直径 MWD (mm) | 大团聚体含量 R0.25 (%) | 几何平均直径 GMD (mm) | 土壤分形维数(Dd) | 根表面积 | 平均直径 | 黏聚力 Cohesion (kPa) |
---|---|---|---|---|---|---|---|
1 | 0.907 | 0.761 | 0.931 | 2.411 | 2.775 | 0.265 | 10.6 |
2 | 0.938 | 0.764 | 0.923 | 2.423 | 3.053 | 0.233 | 10.9 |
3 | 1.101 | 0.831 | 0.988 | 2.322 | 11.642 | 0.354 | 11.1 |
4 | 1.063 | 0.811 | 0.969 | 2.301 | 10.706 | 0.398 | 10.9 |
5 | 0.921 | 0.752 | 0.910 | 2.398 | 3.692 | 0.291 | 10.5 |
6 | 0.930 | 0.723 | 0.975 | 1.805 | 2.221 | 0.261 | 11.2 |
7 | 1.234 | 0.814 | 1.015 | 1.481 | 23.456 | 0.795 | 13.9 |
8 | 1.322 | 0.875 | 1.190 | 1.742 | 24.900 | 0.458 | 14.1 |
9 | 1.059 | 0.822 | 0.985 | 1.728 | 10.846 | 0.361 | 12.6 |
10 | 0.920 | 0.719 | 0.972 | 1.799 | 2.527 | 0.277 | 10.5 |
表2 预测样本
Table 2 Prediction samples
组 | 平均重量直径 MWD (mm) | 大团聚体含量 R0.25 (%) | 几何平均直径 GMD (mm) | 土壤分形维数(Dd) | 根表面积 | 平均直径 | 黏聚力 Cohesion (kPa) |
---|---|---|---|---|---|---|---|
1 | 0.907 | 0.761 | 0.931 | 2.411 | 2.775 | 0.265 | 10.6 |
2 | 0.938 | 0.764 | 0.923 | 2.423 | 3.053 | 0.233 | 10.9 |
3 | 1.101 | 0.831 | 0.988 | 2.322 | 11.642 | 0.354 | 11.1 |
4 | 1.063 | 0.811 | 0.969 | 2.301 | 10.706 | 0.398 | 10.9 |
5 | 0.921 | 0.752 | 0.910 | 2.398 | 3.692 | 0.291 | 10.5 |
6 | 0.930 | 0.723 | 0.975 | 1.805 | 2.221 | 0.261 | 11.2 |
7 | 1.234 | 0.814 | 1.015 | 1.481 | 23.456 | 0.795 | 13.9 |
8 | 1.322 | 0.875 | 1.190 | 1.742 | 24.900 | 0.458 | 14.1 |
9 | 1.059 | 0.822 | 0.985 | 1.728 | 10.846 | 0.361 | 12.6 |
10 | 0.920 | 0.719 | 0.972 | 1.799 | 2.527 | 0.277 | 10.5 |
土样 | 样本数 Sample size | 黏聚力 Cohesion (c, kPa) | 内摩擦角 Internalfriction angle |
---|---|---|---|
含根土Root-bearing soil | 60 | 12.7 | 10.2 |
素土Plain soil | 10 | 9.2 | 10.3 |
表3 直剪试验数据统计
Table 3 Statistical statistics of direct shear test
土样 | 样本数 Sample size | 黏聚力 Cohesion (c, kPa) | 内摩擦角 Internalfriction angle |
---|---|---|---|
含根土Root-bearing soil | 60 | 12.7 | 10.2 |
素土Plain soil | 10 | 9.2 | 10.3 |
处理 Treatment | 土样 | 样本数 | 颗粒与统计 | ||||
---|---|---|---|---|---|---|---|
>2 mm | 2~1 mm | 1~0.5 mm | 0.5~0.25 mm | ≤0.25 mm | |||
干筛 | 含根土 | 30 | 36.84 | 32.67 | 19.07 | 5.81 | 5.61 |
素土 | 30 | 26.44 | 31.27 | 22.11 | 8.90 | 11.28 | |
湿筛 | 含根土Root-bearing soil | 30 | 20.02 | 20.53 | 20.60 | 21.37 | 17.48 |
素土 | 30 | 14.02 | 15.40 | 18.63 | 20.21 | 31.74 |
表4 团聚试验颗粒统计
Table 4 Statistical result of particles in agglomeration test
处理 Treatment | 土样 | 样本数 | 颗粒与统计 | ||||
---|---|---|---|---|---|---|---|
>2 mm | 2~1 mm | 1~0.5 mm | 0.5~0.25 mm | ≤0.25 mm | |||
干筛 | 含根土 | 30 | 36.84 | 32.67 | 19.07 | 5.81 | 5.61 |
素土 | 30 | 26.44 | 31.27 | 22.11 | 8.90 | 11.28 | |
湿筛 | 含根土Root-bearing soil | 30 | 20.02 | 20.53 | 20.60 | 21.37 | 17.48 |
素土 | 30 | 14.02 | 15.40 | 18.63 | 20.21 | 31.74 |
项目 Item | 平均重量直径MWD | 大团聚体 含量R0.25 | 几何平均直径GMD | 土壤分形 维数Dd | 根 | |||
---|---|---|---|---|---|---|---|---|
长度 | 表面积SA | 平均直径 | 体积Volume | |||||
黏聚力Cohesion | 0.414 | 0.380 | 0.301 | -0.146 | 0.422 | 0.395 | 0.164 | 0.174 |
表5 各因素与黏聚力的相关系数
Table 5 Correlation coefficient between each factor and cohesion
项目 Item | 平均重量直径MWD | 大团聚体 含量R0.25 | 几何平均直径GMD | 土壤分形 维数Dd | 根 | |||
---|---|---|---|---|---|---|---|---|
长度 | 表面积SA | 平均直径 | 体积Volume | |||||
黏聚力Cohesion | 0.414 | 0.380 | 0.301 | -0.146 | 0.422 | 0.395 | 0.164 | 0.174 |
样本 | 项目 Item | BP | PSO-BP | FWA-BP | ||||||
---|---|---|---|---|---|---|---|---|---|---|
S=3 | S=4 | S=13 | S=3 | S=4 | S=13 | S=3 | S=4 | S=13 | ||
1 | 期望值 | 10.6 | 10.6 | 10.6 | 10.6 | 10.6 | 10.6 | 10.6 | 10.6 | 10.6 |
预测值 | 12.6 | 12.6 | 11.6 | 11.2 | 11.2 | 11.0 | 11.3 | 10.8 | 10.9 | |
相对误差 | 18.62 | 18.81 | 9.87 | 6.12 | 6.02 | 4.24 | 6.77 | 1.71 | 2.62 | |
2 | 期望值 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 |
预测值 | 12.6 | 12.6 | 11.6 | 11.2 | 11.2 | 11.1 | 11.3 | 11.0 | 10.9 | |
相对误差 | 15.43 | 15.43 | 6.86 | 3.11 | 3.21 | 2.03 | 3.86 | 0.61 | 0.00 | |
3 | 期望值 | 11.1 | 11.1 | 11.1 | 11.1 | 11.1 | 11.1 | 11.1 | 11.1 | 11.1 |
预测值 | 12.5 | 12.6 | 11.8 | 13.0 | 13.0 | 12.8 | 12.7 | 13.2 | 11.6 | |
相对误差 | 13.04 | 13.91 | 6.59 | 17.20 | 17.50 | 15.32 | 14.83 | 18.76 | 4.46 | |
4 | 期望值 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 |
预测值 | 12.5 | 12.6 | 11.7 | 12.6 | 12.5 | 12.5 | 12.1 | 12.7 | 11.3 | |
相对误差 | 15.13 | 15.94 | 7.77 | 15.78 | 14.32 | 14.71 | 11.35 | 16.62 | 3.56 | |
5 | 期望值 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 |
预测值 | 12.6 | 12.6 | 11.7 | 11.3 | 11.2 | 11.5 | 11.3 | 10.9 | 10.7 | |
相对误差 | 19.79 | 19.64 | 11.03 | 7.61 | 6.71 | 9.06 | 7.78 | 4.01 | 2.26 | |
6 | 期望值 | 11.2 | 11.2 | 11.2 | 11.2 | 11.2 | 11.2 | 11.2 | 11.2 | 11.2 |
预测值 | 12.6 | 12.0 | 11.7 | 11.7 | 11.2 | 11.3 | 11.3 | 11.2 | 11.3 | |
相对误差 | 12.70 | 7.16 | 4.16 | 4.39 | 0.30 | 1.00 | 0.98 | 0.00 | 1.13 | |
7 | 期望值 | 13.9 | 13.9 | 13.4 | 13.9 | 13.9 | 13.9 | 13.9 | 13.9 | 13.9 |
预测值 | 12.6 | 12.6 | 13.4 | 13.3 | 13.4 | 13.7 | 13.1 | 13.7 | 13.4 | |
相对误差 | -9.13 | -9.70 | -3.79 | -4.18 | -3.34 | -1.47 | -5.85 | -1.74 | -3.49 | |
8 | 期望值 | 14.1 | 14.1 | 14.1 | 14.1 | 14.1 | 14.1 | 14.1 | 14.1 | 14.1 |
预测值 | 12.6 | 12.7 | 13.4 | 14.3 | 14.3 | 14.1 | 13.5 | 13.9 | 14.3 | |
相对误差 | -10.99 | -10.06 | -4.73 | 1.58 | 1.45 | -0.24 | -4.22 | -1.26 | 1.45 | |
9 | 期望值 | 12.6 | 12.6 | 12.6 | 12.6 | 12.6 | 12.6 | 12.6 | 12.6 | 12.6 |
预测值 | 12.6 | 12.6 | 12.2 | 13.6 | 12.5 | 12.5 | 12.6 | 12.6 | 12.6 | |
相对误差 | 0.00 | 0.00 | -2.84 | 7.67 | -1.05 | -1.05 | 0.00 | 0.00 | 0.00 | |
10 | 期望值 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 |
预测值 | 12.6 | 12.0 | 11.7 | 11.7 | 11.2 | 11.2 | 11.3 | 11.1 | 11.2 | |
相对误差 | 20.18 | 14.20 | 11.12 | 11.46 | 6.90 | 6.99 | 7.70 | 6.12 | 7.44 | |
最大误差 | 20.18 | 19.64 | 11.12 | 17.20 | 17.50 | 9.06 | 14.83 | 18.76 | 7.44 |
表6 不同隐含层神经元预测结果对比
Table 6 Comparison of prediction results of different hidden layer neurons
样本 | 项目 Item | BP | PSO-BP | FWA-BP | ||||||
---|---|---|---|---|---|---|---|---|---|---|
S=3 | S=4 | S=13 | S=3 | S=4 | S=13 | S=3 | S=4 | S=13 | ||
1 | 期望值 | 10.6 | 10.6 | 10.6 | 10.6 | 10.6 | 10.6 | 10.6 | 10.6 | 10.6 |
预测值 | 12.6 | 12.6 | 11.6 | 11.2 | 11.2 | 11.0 | 11.3 | 10.8 | 10.9 | |
相对误差 | 18.62 | 18.81 | 9.87 | 6.12 | 6.02 | 4.24 | 6.77 | 1.71 | 2.62 | |
2 | 期望值 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 |
预测值 | 12.6 | 12.6 | 11.6 | 11.2 | 11.2 | 11.1 | 11.3 | 11.0 | 10.9 | |
相对误差 | 15.43 | 15.43 | 6.86 | 3.11 | 3.21 | 2.03 | 3.86 | 0.61 | 0.00 | |
3 | 期望值 | 11.1 | 11.1 | 11.1 | 11.1 | 11.1 | 11.1 | 11.1 | 11.1 | 11.1 |
预测值 | 12.5 | 12.6 | 11.8 | 13.0 | 13.0 | 12.8 | 12.7 | 13.2 | 11.6 | |
相对误差 | 13.04 | 13.91 | 6.59 | 17.20 | 17.50 | 15.32 | 14.83 | 18.76 | 4.46 | |
4 | 期望值 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 | 10.9 |
预测值 | 12.5 | 12.6 | 11.7 | 12.6 | 12.5 | 12.5 | 12.1 | 12.7 | 11.3 | |
相对误差 | 15.13 | 15.94 | 7.77 | 15.78 | 14.32 | 14.71 | 11.35 | 16.62 | 3.56 | |
5 | 期望值 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 |
预测值 | 12.6 | 12.6 | 11.7 | 11.3 | 11.2 | 11.5 | 11.3 | 10.9 | 10.7 | |
相对误差 | 19.79 | 19.64 | 11.03 | 7.61 | 6.71 | 9.06 | 7.78 | 4.01 | 2.26 | |
6 | 期望值 | 11.2 | 11.2 | 11.2 | 11.2 | 11.2 | 11.2 | 11.2 | 11.2 | 11.2 |
预测值 | 12.6 | 12.0 | 11.7 | 11.7 | 11.2 | 11.3 | 11.3 | 11.2 | 11.3 | |
相对误差 | 12.70 | 7.16 | 4.16 | 4.39 | 0.30 | 1.00 | 0.98 | 0.00 | 1.13 | |
7 | 期望值 | 13.9 | 13.9 | 13.4 | 13.9 | 13.9 | 13.9 | 13.9 | 13.9 | 13.9 |
预测值 | 12.6 | 12.6 | 13.4 | 13.3 | 13.4 | 13.7 | 13.1 | 13.7 | 13.4 | |
相对误差 | -9.13 | -9.70 | -3.79 | -4.18 | -3.34 | -1.47 | -5.85 | -1.74 | -3.49 | |
8 | 期望值 | 14.1 | 14.1 | 14.1 | 14.1 | 14.1 | 14.1 | 14.1 | 14.1 | 14.1 |
预测值 | 12.6 | 12.7 | 13.4 | 14.3 | 14.3 | 14.1 | 13.5 | 13.9 | 14.3 | |
相对误差 | -10.99 | -10.06 | -4.73 | 1.58 | 1.45 | -0.24 | -4.22 | -1.26 | 1.45 | |
9 | 期望值 | 12.6 | 12.6 | 12.6 | 12.6 | 12.6 | 12.6 | 12.6 | 12.6 | 12.6 |
预测值 | 12.6 | 12.6 | 12.2 | 13.6 | 12.5 | 12.5 | 12.6 | 12.6 | 12.6 | |
相对误差 | 0.00 | 0.00 | -2.84 | 7.67 | -1.05 | -1.05 | 0.00 | 0.00 | 0.00 | |
10 | 期望值 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 | 10.5 |
预测值 | 12.6 | 12.0 | 11.7 | 11.7 | 11.2 | 11.2 | 11.3 | 11.1 | 11.2 | |
相对误差 | 20.18 | 14.20 | 11.12 | 11.46 | 6.90 | 6.99 | 7.70 | 6.12 | 7.44 | |
最大误差 | 20.18 | 19.64 | 11.12 | 17.20 | 17.50 | 9.06 | 14.83 | 18.76 | 7.44 |
期望值 | BP | PSO-BP | FWA-BP | |||
---|---|---|---|---|---|---|
预测值 Predicted value (kPa) | 相对误差 | 预测值 Predicted value (kPa) | 相对误差 | 预测值 Predicted value (kPa) | 相对误差 | |
10.6 | 11.6 | 9.87 | 11.0 | 4.24 | 10.9 | 2.62 |
10.9 | 11.6 | 6.86 | 11.1 | 2.03 | 10.9 | 0.41 |
11.1 | 11.8 | 6.59 | 11.8 | 6.32 | 11.6 | 4.46 |
10.9 | 11.7 | 7.77 | 11.5 | 5.54 | 11.3 | 3.56 |
10.5 | 11.7 | 11.03 | 11.5 | 9.06 | 10.7 | 2.26 |
11.2 | 11.7 | 4.16 | 11.3 | 1.00 | 11.3 | 1.13 |
13.9 | 13.4 | -3.79 | 13.7 | -1.47 | 13.4 | -3.49 |
14.1 | 13.4 | -4.73 | 14.1 | -0.24 | 14.3 | 1.45 |
12.6 | 12.2 | -2.84 | 12.5 | -1.05 | 12.6 | -0.24 |
10.5 | 11.7 | 11.12 | 11.2 | 6.99 | 11.3 | 7.44 |
MRE (%) | 11.12 | 9.06 | 7.44 | |||
ARE (%) | 4.60 | 3.24 | 1.96 |
表7 不同神经网络预测值及误差对比
Table 7 Comparison of predicted values and errors of different neural networks
期望值 | BP | PSO-BP | FWA-BP | |||
---|---|---|---|---|---|---|
预测值 Predicted value (kPa) | 相对误差 | 预测值 Predicted value (kPa) | 相对误差 | 预测值 Predicted value (kPa) | 相对误差 | |
10.6 | 11.6 | 9.87 | 11.0 | 4.24 | 10.9 | 2.62 |
10.9 | 11.6 | 6.86 | 11.1 | 2.03 | 10.9 | 0.41 |
11.1 | 11.8 | 6.59 | 11.8 | 6.32 | 11.6 | 4.46 |
10.9 | 11.7 | 7.77 | 11.5 | 5.54 | 11.3 | 3.56 |
10.5 | 11.7 | 11.03 | 11.5 | 9.06 | 10.7 | 2.26 |
11.2 | 11.7 | 4.16 | 11.3 | 1.00 | 11.3 | 1.13 |
13.9 | 13.4 | -3.79 | 13.7 | -1.47 | 13.4 | -3.49 |
14.1 | 13.4 | -4.73 | 14.1 | -0.24 | 14.3 | 1.45 |
12.6 | 12.2 | -2.84 | 12.5 | -1.05 | 12.6 | -0.24 |
10.5 | 11.7 | 11.12 | 11.2 | 6.99 | 11.3 | 7.44 |
MRE (%) | 11.12 | 9.06 | 7.44 | |||
ARE (%) | 4.60 | 3.24 | 1.96 |
模型 Model | 均方根误差 RMSE | 平均绝对误差 MAE | 平均绝对百分比误差 MAPE (%) |
---|---|---|---|
BP | 0.881678 | 0.774666 | 6.7924 |
PSO-BP | 0.893399 | 0.732057 | 6.4062 |
FWA-BP | 0.779353 | 0.625993 | 5.6679 |
表8 不同模型的统计检验
Table 8 Statistical tests of different models
模型 Model | 均方根误差 RMSE | 平均绝对误差 MAE | 平均绝对百分比误差 MAPE (%) |
---|---|---|---|
BP | 0.881678 | 0.774666 | 6.7924 |
PSO-BP | 0.893399 | 0.732057 | 6.4062 |
FWA-BP | 0.779353 | 0.625993 | 5.6679 |
1 | Liu L M, Qiu W M, Xu W N, et al. Discussion on traditional slope protection and ecological slope protection technology. Journal of China Three Gorges University (Natural Sciences), 2007, 99(6): 528-532. |
刘黎明, 邱卫民, 许文年, 等. 传统护坡与生态护坡比较与分析. 三峡大学学报(自然科学版), 2007, 99(6): 528-532. | |
2 | Yang Y H, Wang C H, Liu S Z, et al. Experimental research on improving shear strength of soil in surface landslide by root system of different vegetation type. Research of Soil and Water Conservation, 2007, 61(2): 233-235. |
杨永红, 王成华, 刘淑珍, 等.不同植被类型根系提高浅层滑坡土体抗剪强度的试验研究.水土保持研究, 2007, 61(2): 233-235. | |
3 | Liu X Y, Gui Y, Luo S H, et al. Experimental research on improving shear strength in slope protection by plant root system. Journal of Jiangxi University of Science and Technology, 2013, 34(3): 32-37. |
刘小燕, 桂勇, 罗嗣海, 等. 植物根系固土护坡抗剪强度试验研究.江西理工大学学报, 2013, 34(3): 32-37. | |
4 | Lv J, Gao J R, Wang Y, et al. Effects of different plant revetments on the soil physical feature. Research of Soil and Water Conservation, 2010, 17(3): 101-104, 109. |
吕晶, 高甲荣, 王颖, 等. 不同护坡植物对岸坡土壤性质的影响及效应分析. 水土保持研究, 2010, 17(3): 101-104, 109. | |
5 | Chen C F, Liu H X, Li Y P. Study on grassroots-reinforced soil by laboratory triaxial test. Rock and Soil Mechanics, 2007, 141(10): 2041-2045. |
陈昌富, 刘怀星, 李亚平. 草根加筋土的室内三轴试验研究. 岩土力学, 2007, 141(10): 2041-2045. | |
6 | Lu L X, Zeng B. Enhancement effects of dominant riparian diffuse-rooted plants of Jialing River in Three Gorges reservoir regine on soil anti-scouribility. Journal of Southwest China Normal University (Natural Science Edition), 2006(3): 157-161. |
卢立霞, 曾波. 三峡库区嘉陵江岸生优势须根系植物根系对土壤抗冲性的增强效应研究. 西南师范大学学报(自然科学版), 2006(3): 157-161. | |
7 | Feng G J, Shen F, Li L R, et al. Experimental study on shear strength of root-soil complex. Journal of Southwest China Normal University (Natural Science Edition), 2013, 38(7): 129-133. |
冯国建, 沈凡, 李丽蓉, 等. 根土复合体抗剪强度试验研究. 西南师范大学学报(自然科学版), 2013, 38(7): 129-133. | |
8 | Zuo X F, Wang L, Zheng F L, et al. Effects of freeze-thaw cycles and soil properties on mollisol shear strength in Chinese black soil region. Research of Soil and Water Conservation, 2020, 34(2): 30-35, 42. |
左小锋, 王磊, 郑粉莉, 等. 冻融循环和土壤性质对东北黑土抗剪强度的影响. 水土保持学报, 2020, 34(2): 30-35, 42. | |
9 | Wang R Z, Chen Y, Li T, et al. Root distribution characteristics of Vetiveria zizanioides and Digitaria sanguinalis and their effects on the anti-erodibility of soil in slopelands. Acta Prataculturae Sinica, 2017, 26(7): 45-54. |
王润泽, 谌芸, 李铁, 等. 香根草和马唐的根系特征及对坡地紫色土抗侵蚀性的影响.草业学报, 2017, 26(7): 45-54. | |
10 | Zhang C T, Ma J G, Ding M J, et al. Tensile properties of single root and shear properties of root-soil complex of Bashania qiaojiaensis. Journal of Sichuan Agricultural University, 2022, 40(6): 883-892. |
张春涛, 马建刚, 丁明净, 等. 蔓竹单根抗拉及根-土复合体抗剪特性. 四川农业大学学报, 2022, 40(6): 883-892. | |
11 | Cheng H, Jiang H, Zhang L, et al. Hydraulic separation of the root-soil complex in Festuca arundinacea. Pratacultural Science, 2022, 39(11): 2350-2360. |
程洪, 江辉, 张路, 等.高羊茅根土复合体水力分离研究. 草业科学, 2022, 39(11): 2350-2360. | |
12 | Dang W W,Gao C Z,Dang F N. Research on shear strength index of Xi’an Loess based on improved BP neural network. Journal of Water Resources and Architectural Engineering, 2009, 7(2): 1-4, 13. |
党维维, 高闯洲, 党发宁. 基于改进的BP神经网络对西安黄土抗剪强度指标的研究. 水利与建筑工程学报, 2009, 7(2): 1-4, 13. | |
13 | Jin K H, Yang T, Huo S Y, et al. Research on prediction methods of shear strength of rolled clay based on different PSO-ELM models. Research of Soil and Water Conversation, 2022, 29(3): 213-219, 227. |
金坎辉, 杨涛, 霍树义, 等. 基于不同PSO-ELM模型的碾压黏土抗剪强度预测方法研究. 水土保持研究, 2022, 29(3): 213-219, 227. | |
14 | Jiang W, Ouyang Y, Yan J Z, et al. Inversion iterative correction method for estimating shear strength of rock and soil mass in slope engineering. Rock and Soil Mechanics, 2022, 43(8): 2287-2295. |
江巍, 欧阳晔, 闫金洲, 等. 边坡岩土体抗剪强度的逆向迭代修正反演方法. 岩土力学, 2022, 43(8): 2287-2295. | |
15 | Huang F M, Yin K L, Jiang S H, et al. Landslide susceptibility assessment based on clustering analysis and support vector machine. Chinese Journal of Rock Mechanics and Engineering, 2018, 37(1): 156-167. |
黄发明, 殷坤龙, 蒋水华, 等. 基于聚类分析和支持向量机的滑坡易发性评价. 岩石力学与工程学报, 2018, 37(1): 156-167. | |
16 | Cao Y Y, Su X M, Zhou Z C, et al, Spatial differences in, and factors influencing,the shear strength of typical herb root-soil complexes in the Loess Plateau of China. Acta Prataculturae Sinica, 2023, 32(5): 94-105. |
曹玉莹, 苏雪萌, 周正朝, 等. 黄土高原典型草本植物根-土复合体抗剪性能的空间差异性及其影响因素研究. 草业学报, 2023, 32(5): 94-105. | |
17 | Li T, Wang R Z, Chen Y, et al. Effects of polyacrylamide and grass root system on shear strength and physical properties of purple soil on barren slopes. Acta Prataculturae Sinica, 2018, 27(2): 69-78. |
李铁, 王润泽, 谌芸, 等. PAM和草类根系对荒坡紫色土物理性质与抗剪性能的影响. 草业学报, 2018, 27(2): 69-78. | |
18 | Yang Y C, Mo Y J, Wang Z F, et al. Experimental study on anti-water erosion and shear strength of soil-root composite. Journal of China Agricultural University, 1996(2): 31-38. |
杨亚川, 莫永京, 王芝芳, 等. 土壤-草本植被根系复合体抗水蚀强度与抗剪强度的试验研究. 中国农业大学学报, 1996(2): 31-38. | |
19 | Song L, Zhu H L, Li G R, et al. Comparative Research for two experimental methods to determine soil-root composite shear strength. Research of Soil and Water Conversation, 2016, 23(4): 282-287. |
宋路, 朱海丽, 李国荣, 等. 两种方法测定根-土复合体抗剪强度试验对比研究. 水土保持研究, 2016, 23(4): 282-287. | |
20 | Wang B H, Zhu L Q. Experimental study on shear strength of herbal root-soil composites under different arrangement modes. Journal of Soil and Water Conversation, 2018, 32(6): 118-122. |
王保辉, 朱连奇. 不同布根形式对草本植物根土复合体抗剪强度试验. 水土保持学报, 2018, 32(6): 118-122. | |
21 | Huang Z, Tian F P, Liu Y, et al. Effects of different grassland types on particle size distribution and stability of water stable aggregate on the Loess plateau. Acta Prataculturae Sinica, 2017, 26(11): 216-221. |
黄泽, 田福平, 刘玉, 等. 黄土高原不同草地类型对水稳性团聚体粒径分布及稳定性的影响. 草业学报, 2017, 26(11): 216-221. | |
22 | Luo N, Shu Y G, Chen M J, et al. Soil structure and fractal characteristics of different land categories in a karst rocky desertification area. Acta Prataculturae Sinica, 2020, 29(7): 11-22. |
罗楠, 舒英格, 陈梦军, 等. 喀斯特山区不同草地土壤结构及分形特征. 草业学报, 2020, 29(7): 11-22. | |
23 | Liu W L, Wu J G, Fu M J, et al. Effect of different cultivation years on composition and stability of soil aggregate fractions in orchard. Journal of Soil and Water Conversation, 2014, 28(1): 129-135. |
刘文利, 吴景贵, 傅民杰, 等. 种植年限对果园土壤团聚体分布与稳定性的影响. 水土保持学报, 2014, 28(1): 129-135. | |
24 | Qiu L P, Zhang X C, Zhang J A. Distribution of nutrients and enzymes in Loess Plateau soil aggregates after long-term fertilization. Acta Ecologica Sinica, 2006, 26(2): 364-372. |
邱莉萍, 张兴昌, 张晋爱. 黄土高原长期培肥土壤团聚体中养分和酶的分布. 生态学报, 2006, 26(2): 364-372. | |
25 | Zhang X. Research on traffic prediction of urban road traffic flow date and microblog related to traffic. Tianjin: Tianjin University, 2018. |
张行. 面向城市道路交通流数据与交通微博文本的交通预测研究. 天津: 天津大学, 2018. | |
26 | Tan Y, Shi Y, Tan K C. Fireworks algorithm for optimization. Beijing: Springer-Verlag, 2010: 355-364. |
27 | Najafzadeh M, Barani G A. Comparison of group method of data handling based genetic programming and back propagation systems to predict scour depth around bridge piers. Scientia Iranica, 2011, 18(6): 1207-1213. |
28 | Liu J, Zheng S, Tan Y. Analysis on global convergence and time complexity of fireworks algorithm. Beijing: Institute of Electrical and Electronics Engineers, 2014: 3207-3213. |
29 | Bavel C H M. Mean weight-diameter of soil aggregates as a statistical index of aggregation. Soil Science Society of America Journal, 1950, 14(C): 20-23. |
30 | Dai W Z. A method of multiobjective synthetic evaluation based on artificial neural networks and its applications. Systems Engineering-Theory and Practice, 1999, 19(5): 29-34, 40. |
戴文战. 基于三层BP网络的多指标综合评估方法及应用. 系统工程理论与实践, 1999, 19(5): 29-34, 40. | |
31 | Liu E Q. Classification and detection of household garbage based on deep learning. Taiyuan: Shanxi University, 2020. |
刘恩乾. 基于深度学习的生活垃圾分类和检测. 太原: 山西大学, 2020. | |
32 | He J C. Research on trajectory planning of multi-target manipulator based on practice swarm optimization. Hengyang: University of South China, 2020. |
何建成. 基于粒子群算法多目标机械臂轨迹规划研究. 衡阳: 南华大学, 2020. |
[1] | 魏艳, 刘有斌, 刘枭宏, 谌芸, 颜哲豪, 都艺芝. 紫色土区拉巴豆和紫花苜蓿根-土复合体抗剪性能研究[J]. 草业学报, 2023, 32(8): 82-90. |
[2] | 何伟鹏, 胡夏嵩, 刘昌义, 李璇, 李希来, 付江涛, 卢海静, 杨馥铖, 李国荣. 黄河源区不同禁牧年限对垂穗披碱草单根及其根-土复合体力学强度特征的影响[J]. 草业学报, 2023, 32(5): 106-117. |
[3] | 曹玉莹, 苏雪萌, 周正朝, 郑群威, 岳佳辉. 黄土高原典型草本植物根-土复合体抗剪性能的空间差异性及其影响因素研究[J]. 草业学报, 2023, 32(5): 94-105. |
[4] | 张桐瑞, 李富翠, 李辉, 季双旋, 范志浩, 陈雨峰, 晁跃辉, 韩烈保. 草垫植入对混合草坪坪床稳定性和表观质量的影响[J]. 草业学报, 2020, 29(8): 27-36. |
[5] | 周涛, 谌芸, 王润泽, 李铁, 唐菡, 翟婷婷, 刘枭宏. 种草和施用聚丙烯酰胺对荒坡紫色土抗剪和抗蚀性能的影响研究[J]. 草业学报, 2019, 28(3): 62-73. |
[6] | 李铁, 王润泽, 谌芸, 何丙辉, 周涛, 吴晨, 刘枭宏. PAM和草类根系对荒坡紫色土物理性质与抗剪性能的影响[J]. 草业学报, 2018, 27(2): 69-78. |
[7] | 刘昌义, 胡夏嵩, 窦增宁, 李希来, 徐志闻. 黄河源区高寒草地植被根-土复合体抗剪强度试验及退化程度阈值确定[J]. 草业学报, 2017, 26(9): 14-26. |
[8] | 丁文斌, 何文健, 史东梅, 蒋光毅, 蒋平, 常松果. 干湿作用对紫色土坡耕地生物埂土壤抗剪强度衰减-恢复效应[J]. 草业学报, 2017, 26(6): 56-67. |
[9] | 魏淑珍,王菲凤,张江山. 基于SOFM网络法的天然草地分类[J]. 草业学报, 2011, 20(1): 175-182. |
[10] | 张金屯,聂二保,向春玲. 五台山亚高山草甸的人工神经网络分类与排序[J]. 草业学报, 2009, 18(4): 35-40. |
阅读次数 | ||||||
全文 |
|
|||||
摘要 |
|
|||||