Critical Chain Scheduling and Buffer Management (CC/BM) has shown to provide an effective approach for building robust project schedules and to offer a valuable control tool for coping with schedule variability. Yet, the current buffer monitoring mechanism faces a problem of neglecting the dynamic feature of the project execution and related activity information when taking corrective actions. The schedule risk analysis (SRA) method in a traditional PERT framework, on the other hand, provides important information about the relative activity criticality in relation to the project duration which can highlight management focus. It is implied, however, that control actions are independent from the current project schedule performance. This paper attempts to research these defects of both tracking methods and proposes a new project schedule monitoring framework by introducing the activity cruciality index as a trigger for effective expediting to be integrated into the buffer monitoring process. Furthermore, dynamic action threshold settings that depend on the project progress as well as the buffer penetration are presented and examined in order to exhibit a more accurate control system. Our computational experiment demonstrates the relative dominance of the integrated schedule monitoring methods compared to the predominant buffer management approach in generating better control actions with less effort and an increased tracking efficiency, especially when the increasing buffer trigger point is combined with decreasing sensitivity action threshold values.

No abstract

No abstract

No abstract

The concept of feeding buffer, proposed by Goldratt (1997), has attracted much attention to the world. There are many researches on the feeding buffer size. However, actually in real practice, even the same project may need to insert with different size of feeding buffer as different team with different level of skilled workers and different project environment with different kind of uncertainty, resulting in different kind of safety time (Newbold, 2008). In this paper, we investigate the impact of feeding buffer size, service level (deadline), and uncertainty of activity duration on the timely project completion probability in a stochastic resource-constrained project scheduling (RCPSP) environment. The results indicate that there has a huge interaction between feeding buffer, deadline, and different level of uncertainty in the project activity duration. Extensive computational results will be presented on benchmark library PSBLIB instances and statistic analysis will be conducted by using SAS PROC GLM.