The Research On Duration Prediction Based On PSO-SVM For Black And Odorous Water Bodies Renovating Engineering Project In Y City

Over the 41 years of Economic Reform and open up,urbanization developed rapidly,however,the pressure on urban ecological environment continues to increase and the problem of water pollution is increasingly prominent.From the 2015 Action Plan for Prevention and Control of Water Pollution to the 14th Five-Year Plan,the remediation of urban black and odorous water bodies has been included in the national ecological governance goal.According to the statistical results of black and odorous water bodies renovating project in Y City of 2021,up to 55%of the projects were delayed,one of the essential reasons is the lack of reliable basis for determining the construction duration.Consequently,it is necessary to provide a scientific,effective,accurate and applicable prediction method for the determination of the construction period of the proposed black and odorous water renovating project in Y City,so as to ensure that the plan of the renovating engineering is reasonable and the overall objective is orderly achieved.Compared with the traditional infrastructure construction,the black and odor water renovating project has the characteristics of late start,less project experience and strong regionality.The common prediction methods are difficult to achieve objective and effective duration prediction.Therefore,a Support Vector Machine(SVM)duration prediction model optimized by Particle Swarm Optimization(PSO)was proposed.Taking Y City as an example,using the 11 completed black and odorous water renovating projects,the study identified 37 factors from the perspective of certainty and uncertainty.The mean of impact value algorithm was used to screen the factors,and 17 influencing factors of the construction period of the black and odorous water renovating project in Y City were obtained.Aiming at the inconvenient building of traditional support vector machine model and the difficulty of parameter determination,PSO-SVM duration prediction model was constructed by introducing support vector machine toolbox and particle swarm optimization algorithm.The selected project samples were used to train and test the model,and the mean squared error and squared correlation coefficient meet requirements.Finally,the trained model was applied to predict the reasonable and ideal construction period of a black and odor water renovating project in Y City.Compared with the traditional duration prediction method,the construction period prediction model based on machine learning algorithm is more scientific and reasonable,reducing the reliance on experts and subjective randomness in the duration prediction of black and odor water renovating project.The prediction results show that the reasonable construction period of the project is 97 days,and the ideal construction period is 107 days,which proves the operability of PSO-SVM construction duration prediction model.The purpose is to provide an objective and reliable prediction method for the construction unit to determine the project duration.