UNDERGROUND PIPELINE GALLERY CONSTRUCTION WASTE PREDICTION BASED ON DBN IN PREVENTION AND CONTROL OF ENVIRONMENTAL POLLUTION

摘要

In order to solve the problems that are difficult to deal with by traditional prediction methods, such as non-linear, time-series and random, and strong di-versity of micro correlation factors, a prediction method of construction waste in urban underground pipe gallery based on Deep Belief Network(DBN) is proposed. Firstly, the Nadam momentum optimiza-tion algorithm is used to train the DBN and obtain the best DBN parameters,a learmning framework for construction waste prediction is formed. Secondly, PLSR was used to replace gradient fine-tuning method in conventional DBN for improving predic-tion accuracy. Meanwhile, a Lyapunov function was constructed to prove convergence of the proposed method in the learing process. Finally, the proposed method is applied to underground pipeline gallery construction waste prediction. The experimental re-sults show that the method has a fast convergence rate and a high prediction accuracy, which can meet the demands for waste prediction.