Municipal Solid Waste Volume Forecasting Based On LSTM In Tianjin

With the rapid development of economy and the acceleration of urbanization process,the migrant population rapidly gathers to the city,resulting in the rapid growth of municipal solid waste.Since 2012,the municipal solid waste in tianjin has been growing at a rate of 10%～15% every year.Therefore,the pressure of cleaning and transportation has gradually increased,and corresponding waste treatment facilities are needed.The urban management department needs to plan the domestic waste treatment facilities according to the estimated municipal solid waste.If the predicted quantity is too low,the treatment facilities may not be able to absorb and transport the municipal solid waste,causing environmental pollution,reducing the quality of life of residents,and aggravating the risk of garbage siege.If the forecast volume is too high,it may lead to idle equipment and waste a lot of resources.Therefore,the accurate prediction of municipal solid waste is of great significance to the rational planning of waste treatment facilities and the improvement of management efficiency of urban management departments.Firstly,this paper analyzed the situation of municipal solid waste in tianjin,sorted out and analyzed the capacity of municipal solid waste harmless treatment in tianjin from the number of harmless waste treatment plants,the rate of harmless treatment and the number of sanitation workers.After reading the literature and analyzing the influencing factors in combination with the situation of tianjin,seven indexes were finally selected as the input indexes of the municipal solid waste prediction model of tianjin.Secondly,based on the current situation that deep learning methods are rarely used in the prediction of municipal solid waste,this paper selects three types of deep learning models of LSTM,GRU and ANN,and constructs four different architectures for training and testing respectively to verify the prediction effect of each model.By comparing the optimal models of different categories,the experimental results show that the MAPE of the LSTM model is 9.27%,with the smallest prediction error.Therefore,the LSTM model is selected as the prediction model of tianjin municipal solid waste in this paper.In addition,this paper predicts the municipal solid waste in tianjin based on different scenarios,and the results show that in 2023,it will float between 3.4196 million tons and 4.7282 million tons.Finally,this paper gives the optimization strategy of household garbage management from three aspects: strengthening garbage treatment capacity,promoting garbage classification,and strengthening garbage collection,storage and transportation system,for the reference of the management department.