Title:prediction Of Airport Surface Subsidence Based On Multi-objective Particle Swarm Optimization Improved Grey Model

Grey model plays an important role in grey system theory,which is suitable for solving the prediction problem of "small sample" and "poor information",and has been widely used in many fields.Settlement prediction is an important research direction,and accurate settlement prediction data can provide reliable decision basis.There are a lot of literatures on settlement prediction,but there are still some deficiencies.Although the prediction of middle and late settlement is suitable for buffer operator theory of grey model,there are few researches on this aspect.In addition,the structure of the grey model itself has defects and needs to be improved.Model improvement generally involves parameter optimization,and overfitting problems exist in common optimization methods.In view of this,this paper studies the shortcomings of gray model and settlement prediction.The main research contents include:(1)Based on the buffer operator theory,a single and double weight power function weakening buffer operator is proposed to eliminate the impact disturbance on data.The deviation of the background value of the grey model is analyzed,and the variable weight background value is constructed according to the integral mean value theorem.The improved grey model(variable weight buffer grey model)is built by combining with the new operator.(2)It is proved that the data sequence with the same symbol is multiplied by a non-zero constant to obtain a new data sequence.Based on the old and new data series,the GM(1,1)model with optimized background value and initial value is established respectively.When the time parameters of the two models are the same,and the relationship between the background values and the background values,and the relationship between the initial values and the initial values are the same as the relationship between the two data series,the accuracy of the two models is the same.(3)The optimal parameter determination method of grey model with parameters is studied,and propose a method of using multi-objective particle swarm optimization to optimize the parameters of the variable weight buffer grey model.The objective function of the algorithm is to minimize the average relative error and the negative value of the correlation degree.The criteria for selecting the optimal parameters of the model are given,which solves the overfitting problem when the single objective algorithm optimizes the parameters of the model.(4)The middle and late sedimentation data is a kind of impact disturbance system,which is suitable for establishing variable weight buffer grey model.A variable weight buffer grey model optimized by multi-objective particle swarm optimization was established on MATLAB 2020a,and successfully applied to the prediction of surface subsidence in the middle and later stages of a certain airport,verifying the effectiveness and superiority of the new method.