Research Of Prediction Of Atmospheric Environmental Pollutants' Concentrations Based On IGA-PSOW-BP Model

With the rapid development of China's society and economy and due to inadequate understanding to negative influence evaluation in industrialization process,a severe environmental pollution has arisen.The environmental pollution includes water pollution,earth pollution,noise pollution and air pollution and so on.And the air pollution is one of the most severe environmental pollutions and has great influence in human's living and producing activities.The air pollution makes great inconvenience in people's life,ruins people's health,causes major losses in economy development which up to billions of dollars every year.In order to bring the air pollution under control,cut down even eliminate the influence of air pollution in human's living and producing activities and society economy development,our country has passed relevant laws and regulations and announced some advanced pollution prevention technologies and other practice to promote comprehensive improvement and protection.Laws and regulations,such as Law of the People's Republic of China on the Prevention and Control of Atmospheric Pollution and others laws and regulations in specific fields about prevention and control of atmospheric pollution,are published.On the prevention and control of atmospheric pollution,prediction of atmospheric environmental pollutants' concentrations can provide foundations for relevant government officers and company decision-makers.This paper employs atmospheric pollution data and meteorological data which have been stored,back propagation feedforward neural network(BP),genetic algorithm(GA)and Particle Swarm Optimization(PSO)algorithm and other algorithms to build complete prediction model of atmospheric environmental pollutants' concentrations to predict each pollutant's concentration.Firstly,this model adopts data preprocessing methods which include abnormal data processing,absent data processing and normalization processing to data which has been stored to make sure every data is complete and valid,and then uses Pearson correlation coefficient to analyse data dependency between different pollutant factors.After which the model chooses one pollutant factors from these relevant factors to make up of the model's input with meteorological data.Using improved genetic algorithm which is called IGA and improved particle swarm optimization algorithm which is called PSO-W and selected pollutant factors and meteorological data to train the model is the last procedure to build this model completely.After trainning,importing the input to the model and the model calculates the prediction results as outputs.Also,in order to check this model's prediction accuracy,this paper does some comparative experiments.After comparing this model with BP model,GA-BP model and PSO-BP model,this experiment proves that the model this paper proposes has excellent prediction accuracy and low time-consuming,outputs accrurate results within a short time and describes the atmospheric pollutants' concentration in the following future.So the model this paper proposes can provides decision basis for government officers to make relevant laws and regulations and enterprise decision-makers to adjust producing activities even ordinary people to plan their lives.