Buffer Monitoring of Critical Chain Projects Based on Support Vector Machine Prediction

Abstract

Uncertainties have a negative impact on the duration of project activities. When an activity faces the higher uncertainty, it is likely to experience larger fluctuations in its duration. This increases the risk of delays. However, classical buffer monitoring methods usually adopt the setting mode of uniform and fixed monitoring time points for different activities, failing to account for differences in uncertainty levels between them, which reduces the effectiveness of project schedule control. Therefore, we propose a dynamic buffer monitoring method combining buffer monitoring and forecasting. Firstly, a duration prediction model based on support vector machine is established to predict the duration of the subsequent activity relying on the duration data of completed activities. Secondly, the buffer consumption rate is calculated according to the predicted activity duration and the corresponding monitoring frequency is obtained. Matlab is finally utilized to verify the method proposed in this paper. The results show that compared with classical buffer monitoring methods, the proposed method achieves the dual optimization of project duration and cost.