基于自私兽群算法优化多尺度熵的区域降水复杂性分析

Analysis of the Complexity of Regional Precipitation Based on the Optimized Multi-scale Entropy Obtained by Selfish Herd Algorithm

为精确识别区域降水的复杂和不确定性特征,探索对于区域降水复杂性的可能影响因素,以北大荒农垦集团有限公司建三江分公司下辖的15个农场为研究区域,采用自私兽群算法对多尺度熵参数进行寻优,利用优化后的多尺度熵对1997—2018年各个农场的月降水复杂性测度进行计算分析,采用ArcGIS软件对月降水复杂性测度进行空间化展示,探索导致降水复杂性的可能原因。结果表明:研究区中部农场的降水复杂性测度等级最高,东北部农场的次之,西南部农场的复杂性测度等级最低。通过分析发现,人类因素对降水复杂性的影响高于自然因素,人类规律性改造自然的过程会降低降水的复杂性。在模型性能上,自私兽群优化算法的寻优效率得到了显著提升。

In order to accurately identify the characteristics of regional precipitation complexity,explore possible factors affecting regional precipitation complexity,Taking the 15 farms under the construction of the Jiansanjiang branch of China Beidahuang Agricultural Reclamation Group Co.,Ltd.as the research area,selfish herd algorithm is used to optimize the multiscale entropy parameters,optimized multiscale entropy is used to calculate and analyze the monthly precipitation complexity measures of each farm from 1997 to 2018,ArcGIS software is used to spatially display the complexity measures,and the possible causes of precipitation complexity are explored.The results show that the farms in the central region have the highest level of precipitation complexity measurement,followed by farms in the northeast,and the farms in the southwest have the lowest level of complexity measurement.Through analysis,it is found that the influence of human factors on the complexity of precipitation in Jiansanjiang area is higher than that of natural factors,and that the process of humans transforming nature regularly will reduce the complexity of precipitation.In terms of model performance,the optimization efficiency of the selfish herd algorithm has been significantly improved.