GM(1,1)改进模型的在智慧水务中的研究与应用

Research and Application of GM(1,1) Improved Model in Smart WaterZhang Yihang1 , Qian Xiaoqun2, Peng Hongyu3

摘要:

智慧水务是智慧城市的重要组成部分，水资源合理的分配调度更是智慧水务系统中的重中之重，如何准确快速的预测未来某时的指标数据在调度过程中尤为关键。通过对智慧水务的供水模型进行了研究，提出了预测模型的应用方案，验证了其在工程中的合理性。在预测模型的研究中,首先采用了经典的均值GM(1,1)模型对水流量序列进行预测，并在此基础上引入马尔科夫链，使用单步和多步加权分别对GM(1,1)模型的误差进行分析和预测，对其结果进行了矫正，最后使用MATLAB对三种方式进行了计算和仿真。比较结果显示马尔科夫链与GM(1,1)模型的结合比单纯的GM(1,1)模型在预测精度上有较大提升，而加权之后马尔科夫的矫正效果比单步的矫正效果更好。

Abstract:

Smart water is an important part of smart cities. The rational allocation of water resources is one of the most significant step in smart water systems. How to predict the future index data accurately and quickly is particularly critical in the scheduling process. Through the study of the smart water supply model, the application scheme of the forecasting model is put forward, and the rationality of the forecasting model in the project is verified. In the study of forecasting model, the classical mean GM (1,1) model is used to predict the water flow series firstly. On this basis, Markov chain is introduced to analyze and predict the errors of GM (1,1) model by using single-step and multi-step weighting, and the results are corrected. Finally, MATLAB was used to calculate and simulate the three methods. The comparison results show that the combination of Markov chain and GM (1,1) model has a greater improvement in prediction accuracy than the simple GM (1,1) model. The Markov chain after weighting is also improved in accuracy compared to the single-step prediction state.