KFFM-based Large-scale Project Earned Value Management Completion Prediction

Abstract

Earned value management lacks research on the project completion trend forecasting method at the following two points: The project completion trend forecast is based on the state indicators at the project monitoring point, and the project execution efficiency before the monitoring time point is rarely considered; the project performance evaluation index How to reduce the non-efficiency noise caused by uncontrollable factors in the existing prediction methods when singularities occur. The above issues will affect the accuracy and stability of the earned value management forecasting system. This paper attempts to introduce the Kalman filtering method (KFFM) into the earned value management completion prediction calculation, and comprehensively considers the project execution efficiency in the logical order of “prediction-measure-correction”, and improves the prediction accuracy through dynamic correction of cost/progress indicators. And effectively reduce the impact of the earned value management forecasting system due to singularities