The municipal solid waste generation distribution prediction system based on FIG–GA-SVR model

Abstract

With the rapid development of economy, the process of urbanization in China has become more and more rapid, and the municipal solid waste (MSW) generation has also increased year by year. The MSW generation prediction is an important prerequisite for MSW management. Considering the uncertainty of MSW generation and the unbalanced economic development in different districts, first, the fuzzy information granulation (FIG) method is used to granulate and predict the three explanatory variables (annual disposable income per capita, GDP and total retail sales of consumer goods) in different districts. Second, the prediction results are substituted into the support vector regression model optimized by genetic algorithm (GA-SVR) to predict the MSW generation per capita. Then, ARIMA model is applied to predict population of each community. After that, the predicted MSW generation per capita is combined with the predicted population of each community to obtain the MSW generation distribution. Finally, the Kriging interpolation method is used to present the MSW generation distribution. The MSW generation distribution prediction in Huangshi, a city of Hubei province in China, is chosen as the case to test the feasibility and effect of the model. The prediction results of Huangshi presented as interval value suggest that the accuracy of MSW distribution prediction can meet the identification requirements of MSW management system. Therefore, the FIG–GA-SVR prediction model proposed in this paper is suitable for interval prediction and has good generalization ability. This model can not only be applied to the prediction of MSW generation, but also can be applied for prediction in other fields.