Research On Construction Dynamic Schedule Control Based On BP Neural Network Prediction And Fuzzy Earned Value Method

Schedule,cost and quality are the main factors that determine the success or failure of a project.Facing the increasingly competitive construction market,contractors must not only control the construction quality of the project,but also strictly control the cost and schedule of the project to enhance their own strength and gain an advantage in the competitive market.Due to the long construction period of engineering projects,managers are often unable to accurately predict the various situations that may occur during the entire process.Therefore,it is necessary to continuously adjust during the project implementation process to dynamically control the entire progress of the project.In addition,due to the complexity of the definition of project objectives and the construction environment,uncertain factors frequently occur during the implementation process,and it may not be possible to express the project completion period deterministically,which has a great impact on the construction schedule control.This article focuses on the research of dynamic schedule control in engineering construction.Traditional Earned Value Management(EVM)methods have many drawbacks when applied to the process of engineering project schedule control.First,the traditional EVM’s progress evaluation indicators are all based on cost considerations.When applied to schedule control,the progress indicators measured in monetary units often cannot intuitively reflect the project progress.Therefore,time-based units should be used in combination.The measured progress index controls the construction progress in real time.Second,the traditional EVM method is calculated based on deterministic values.Due to the influence of the project environment and frequent uncertain factors,in the actual construction process,the project completion period may not be accurately expressed,so it is impossible to accurately obtain deterministic earned value data.Thirdly,when using EVM to control the progress of engineering projects,calculations are based on historical statistics of earned value parameter indicators,and this process is often static.This process does not take into account the impact of future resource planning on the project schedule,nor can it achieve dynamic control of the schedule based on future resource planning adjustments.Therefore,when carrying out dynamic schedule control of engineering construction,it should be processed and improved on the existing EVM method.Based on the problems faced by the current schedule control,this article establishes an innovative construction dynamic schedule control model:based on the combination of BP neural network prediction and fuzzy-earned value management(Fuzzy-Earned Value Management,F-EVM).This dynamic control model of construction schedule is mainly implemented through three modules.The main research work is as follows:(1)Real-time monitoring module based on F-EVM.In order to achieve better project construction schedule control,based on a large number of literature readings,three earned value methods currently used for project schedule control are finally determined:Planned Value Management(PVM)and earned schedule management method(Earned Schedule Management,ESM)and Earned Duration Management(EDM).Since the completion date cannot be expressed deterministically in actual projects,fuzzy theory is introduced and combined with the earned value management method to form the fuzzy earned value method,which realizes real-time monitoring of the project progress during construction.It is worth noting that the fuzzy earned schedule management method has not yet been developed,which is also an innovation in the method of this article.(2)Future EVM parameter prediction module based on BP neural network.When monitoring the construction of the project,if there is a progress warning of the project,it will be transferred to the prediction module for analysis.By using the BP neural network prediction model,the future resource plan is used as the input indicator to predict the EVM-related progress parameters at the future time,so as to obtain the fuzzy earned value schedule performance indicator in the future time,and use it as the basis for dynamic schedule control decisions.(3)Dynamic schedule control decision-making module.In this module,first defuzzified the forecast-based future fuzzy schedule performance indicators,and make a final decision analysis based on the principle of schedule control decisionmaking.When there is a risk of delay in the project,the future resource planning should be adopted Make adjustments to make it meet the schedule control target,to achieve dynamic schedule control for the construction of the project.Finally,through an actual construction case,the dynamic schedule control model proposed in the article was verified and analyzed.The results show that the model can effectively improve the accuracy of schedule control and realize effective prediction of future schedule.This provides a more accurate reference basis for project managers to conduct dynamic schedule management,and is of profound significance to the development of project dynamic control systems.