Study On The Model Of MSW Delivering Quantity Forecasting In Hefei Based On Bp Neural Network Optimized By IABC Algorithm

In recent years,economic development has brought continuous improvement of living standards and consumption structure is constantly upgraded,which makes the daily living garbage increasing,especially in some super cities,the problem of waste removal and landfill space occupation has become a social problem that pollutes urban health and environment,and affects the production and life of citizens.Under the background of the state's vigorously promoting of waste incineration for power generation and the construction of waste incineration equipment and facilities according to the classification of solid waste delivering quantity,we should do well in whole process management of the Municipal Solid Waste even more.Its management should not only be timely and efficient,but also avoid secondary pollution to the living environment.We need to make more accurate predictions of the Municipal Solid Waste delivering quantity,it's quantitative base.The prediction can also provide reference for the reduction of the total waste and the planning and layout of urban environmental sanitation facilities.Most of the existing forecasting methods adopt the growth rate forecasting method and the linear regression forecasting method,and these models are mostly static,so it is often difficult to get high precision forecasting results.This paper introduces the artificial neural network algorithm.BP neural network algorithm has a good nonlinear fitting ability,has been widely used in dealing with nonlinear problems,can effectively solve all kinds of practical problems.However,when the neural network is initialized,its weights and thresholds are randomly assigned by the system,and there are uncertain factors.Therefore,in this paper,artificial bee colony algorithm is used to optimize the weight and threshold of BP algorithm before the model is constructed.Artificial bee colony algorithm is a swarm intelligence algorithm,which simulates the division of foraging behavior of bee colonies.The algorithm finds the optimal solution of the group by comparing the advantages and disadvantages of various selection values and local optimization of individuals.However,artificial bee colony algorithm may fall into local optimal when calculating.In view of this deficiency,an improved bee colony algorithm is proposed in this paper,combining BP algorithm to form Back Propagation Neural Network model Optimized by Improved Artificial Bee Colony Algorithm.In terms of data acquisition and processing,this paper firstly classifies and specifically analyzes the factors that effect the Municipal Solid Waste delivering quantity,factors easy to be quantified as dependent variables are choosen,conducts grey relational degree analysis on them,then the important relational elements with high correlation degree were screened out and substituted into the feedforward neural network model based on MATLAB software.Taking Hefei city as an example,four years of the Municipal Solid Waste delivering quantity from 2019 to 2022 will be 1.685 million tons,1.693 million tons,1.7025 million tons and 1.7153 million tons respectively.Through comparison,it is found that the fitting and prediction results using the optimized model have higher accuracy and smaller error.