Research Of Chengdu Domestic Waste Output Forecast Based On GM-ARIMA Model

Under the background of the rapid development of urbanization in citis, we saw more and more phenomenon of "Besieged by Garbage" in the medium cities.City life garbage and Industrial solid wastes gradually affect the urban sanitation and the human health.The accurate prediction of City life garbage output can provide gist for the sanitation department to make the urban environmental sanitation planning.It can also help the city managers to built city life garbage disposal facilities.Based on this, this thesis has carried out the following work:1. Collection of domestic and foreign research data about city life trash output prediction, choosing models related to predict the output of municipal solid waste (MSW); 2. Analysis the principle of solid waste output forecast model, put forward the optimizationof model; 3. Combined with the model optimization method and establish the corresponding model, the Chengdu city life trash output prediction, and puts forward the control and reduce the corresponding measures and proposals of Chengdu city life trash output.With the work of integrated papers,the following conclusions can be obtained:First, the paper studied the research status of domestic waste output prediction at home and abroad.Using time series model (ARIMA (p, d, q)) and the gray model (GM (1, 1)).From the perspective of model optimization method, the paper put forward optimization method of two kinds of model.It is the parameters optimization method and residual error correction method. In terms of parameter optimization, paper using the intelligent optimization algorithm of particle swarm optimization algorithm (PSO) to optimize the model’s parameter.Established the PSO-GM model based on PSO algorithm; And the residual error correction method is used to optimize the ARIMA model. The results show that two kinds of optimization model put out the precise results.So the model are improved compared to the original model.Second, based on the research of the combination model theory, paper combined the PSO-GM model and residual error correction model of ARIMA (3,1,2).Respectively, set up a simple average combination model, the weighted average of the combined model, and based on particle swarm optimization algorithm (PSO) GM-combination of ARIMA model. Using the data of living garbage processing in January 2013-December 2015 monthly in Chengdu, compaing the three combination model, single PSO-GM model and residual error correction model of ARIMA (3,1,2). Through the following indicators:the mean square error (MSE), mean absolute error (MAE) and mean absolute percentage error (MAPE) to evaluate model precision.The results of the study found that the combined model GM-ARIMA model based on PSO algorithm of prediction accuracy is better than other four kinds of models.Third, choosing the GM-ARIMA model based on the PSO algorithm to make the forecast of leaving garbage production in Chengdu in 2016-2020.Forecast results to the university and Chengdu city life garbage processing of 2.083101 million tons in 2020. Combination forecast results, puts forward some referential suggestions of garbage control and treatmenthe in Chengdu in the future.