A Study On Application Of GA-PSO-BP Neural Network In Prediction Of Air Pollutants Concentration

With increasing degree of industrialization, the rise while promoting economicprosperity, but also caused us serious environmental pollution problems. In recent years,serious frequent fog and haze, give me health and daily lives of the severely affectedaroused concern and worry.In recent years, the state attaches great importance to the prevention and control of airpollution, China’s environmental protection department to respond positively to the call ofnational policy, a comprehensive upgrading of the National urban ambient air monitoringsystem. With the increase of system upgrades and monitoring sites, the level of monitoringtechnology has been greatly improved. In order to more fully understand and grasp thetrend of atmospheric pollutants, provide more comprehensive and timely information to airpollution prevention and control work, to carry out research work to predict atmosphericpollutants is very important. Environmental forecasting of atmospheric pollutants afterdecades of development, forecasting methods and forecasting techniques have been fullyimproved. However, how to improve the prediction accuracy of prediction of air pollutantsis critical.In this paper, based on neural network technology, design and implement a hybridgenetic algorithm and particle swarm algorithm BP network model for predicting airpollutant concentrations. The model based on BP neural network, the introduction of ahybrid genetic algorithm and particle swarm algorithm constructed to achieve the initialweights of BP network optimization, to effectively improve the generalization ability ofneural network, avoiding premature convergence in neural networks local extreme point.The model uses concentration SO2, NO2and PM10conducted experiments.Experimentalresults show that the prediction model has good predictive accuracy, to achieve the desiredeffect.