基于小世界复杂网络的水文诊断模型仿真  

Simulation of Hydrological Diagnosis Model Based on Small World Complex Network

作　　者：裴源博 赵雪花[1] PEI Yuan-bo;ZHAO Xue-hua(Taiyuan University of Technology,College of Water Resources Science and Engineering,Taiyuan Shanxi 030024,China)

机构地区：[1]太原理工大学水利科学与工程学院,山西太原030024

出　　处：《计算机仿真》2022年第4期477-481,共5页Computer Simulation

基　　金：国家重点研发计划(2019YFC0408601);山西省重点研发计划项目(201903D321052);山西省自然科学基金(201901D111060)。

摘　　要：为便于准确预测未来水文信息,提出基于复杂性理论的非一致性水文序列突变诊断方法。分析水文突变驱动因子,包括自然变化、气候变化与人类活动等,从概率分布、均值、方差与变异系数多方面表现水文序列突变统计特征。结合复杂性理论,将诊断模型抽象为一个复杂网络,分别计算最短路径、集聚系数与节点介数,构建一个介于规则和随机之间的小世界复杂网络模型。采用最小二乘算法确定模型目标函数,设置约束条件,获得最优分段线性方程,考虑变异点存在的连续性特征,计算变异点置信的上、下限,更加准确地判断变异点位置,实现水文序列突变诊断。仿真结果证明,上述方法诊断结果与历史资料具有高度相似性,可为后续水文系统分析提供决策依据。A catastrophe diagnosis method of inconsistent hydrological series based on complexity theory is reported for accurately and conveniently predicting hydrological information. The driving factors of the hydrological catastrophe were analyzed in detail(natural change, climate change and human activities, etc.). The statistical characteristics of hydrological series catastrophe included probability distribution, mean value, variance and coefficient of variation. Complexity theory was introduced to abstract the diagnosis model and turn it into a complex network. The shortest path, agglomeration coefficient and node intermediate number were calculated respectively to build a small world complex network model between regular and random. According to the least square method, the objective function of the model was determined, the constraints were set, and the optimal piecewise linear equation was obtained. Based on the continuity characteristics of variation points, the upper and lower limits of confidence were used to judge the location of variation points, thus achieving the mutation diagnosis of hydrological series. The simulation results show that the diagnosis results of this method are consistent with historical data, which can provide a basis for subsequent hydrological system analysis.

关 键 词：复杂性理论 非一致性 水文序列 突变诊断 最小二乘算法 

分 类 号：TV882[水利工程—水利水电工程]