Abstract: Project delay and cost overrun have been common in all industries nowadays. In order to solve these problems, an effective method—critical chain project management (CCPM) is proposed based on the theory of constraints. It is well known that managers will add a large amount of safety time to the project plan. However, this safety time will be consumed during the implementation of the project owing to various factors, such as the student syndrome, which contribute to the project being delayed and increase project costs. CCPM accordingly adopts the concept of buffer to absorb the uncertainties. As one of the core aspects of buffer management, buffer monitoring plays an important role in ensuring the on-time delivery and providing the project with sufficient protection. Although CCPM has improved the efficiency of project management, project duration delay and cost waste still exist due to the unreasonable buffer monitoring procedure. To deal with the problem that the monitoring focus of existing methods is not prominent and the decision information is not comprehensive in the process of buffer monitoring, an integrated buffer monitoring method based on the important characteristics of the activity is proposed in this paper. First, the new method introduces the activity importance information in the project schedule risk analysis method as a basis for identifying those critical activities that needs more attention and stronger actions. The activity importance of the project is measured by comprehensively considering the variability of the activity's duration, the impact of the activity's duration on the completion of the project, and the network location of the activity. And the judgment threshold of the importance index is dynamically set according to the fluctuation degree of the activity duration. Then, the subsequent trend of buffer consumption was forecasted quantitatively based on gray neural network. Considering the project buffer consumption and project follow-up information comprehensively, the degree of deviation of the buffer consumption state is described by the membership function, corrective actions are differentiated for the activity according to the buffer deviation degree. Meanwhile, the buffer monitoring areas and the action threshold of buffer consumption are set during the dynamic project execution. Finally, the computational experiment based on YS company’s smart party building software development project is carried out to validate our model. The results show that the proposed method can control the progress of the project comprehensively and systematically compared with the traditional monitoring methods. It has better performance in indicators such as monitoring the regional frequency distribution, buffer consumption rate, actual project duration and cost performance, thus realizes the dynamic feedback between project planning and schedule adjustment and improves the efficiency of buffer management. Our research will help scholars better to understand the essence of activity criticality and subsequent trend information in buffer monitoring process, and provide guiding strategies for enterprises to improve the performance of project schedule management and ensure the on-time delivery of projects.

Key words: critical chain; buffer management; buffer monitoring and control; buffer subsequent trend information; activity criticality