基于DEA算法的矿山修复生态效益评估模型

DEA algorithm-based model for evaluation of ecological benefit of mine restoration

为准确展现矿山修复生态效益成果,设计基于DEA算法的矿山修复生态效益评估模型。依据矿山地理区域特征和其修复生态效益目的,利用层次分析法,建立矿山修复生态效益评估指标体系;根据获取的一级、二级和三级评估指标,利用DEA算法建立DEA评估模型,通过求解DEA评估模型内的C2R线性规划模型,对决策单元进行判定,当判定为无效时,利用投影定理在保证输出数值一定的情况下降低原来的输入,或在保证输入数值一定的情况下提升输出数值,从而完成矿山修复生态效益评估。实验结果表明:该模型对矿区的修复生态效益指标评估结果与专家评分结果相差不大,且与之相比更加精确,具有较高的评估有效性;评估均方根误差仅为2.78,评估准确性高;模型ROC曲线较为平滑,且灵敏度数值较高,评估稳定性卓越。

In order to accurately exhibit the ecological benefits of mine restoration, a DEA algorithmn-based model was designed for evaluation of ecological benefits of mine restoration. According to the geographical characteristics of the mine and the purpose of the ecological benefits of the mine restoration, the evaluation index system of the ecological benefits of the mine restoration is established by using the analytic hierarchy process. According to the obtained first-level, second-level and third-level evaluation indexes, the DEA algorithm is used to establish the DEA evaluation model. By solving the C2R linear programming model in the DEA evaluation model, a judgment is made on the decision unit. When the judgment is invalid,the projection theorem is used to reduce the original input under the condition that the output value is fixed. Or improve the output value when the input value is fixed, so as to complete the ecological benefit assessment of mine restoration. The experimental results show that the model is more accurate and effective compared with the evaluation results of experts. The root mean square error of evaluation was only 2.78, indicating high accuracy. The ROC curve of the model is smooth, and the sensitivity value is high, and the evaluation stability is excellent.