Buffer sizing approach with dependence assumption between activities in critical chain scheduling

Abstract

The critical chain/buffer management (CC/BM) methodology, proposed by E.M. Goldratt [1997. Critical chain. New York: The North River Press], introduced the concept of buffers to protect the critical chain against the disruptions. The buffer sizes reflect the uncertainty in the estimation of project duration and affect the project scheduling performance. However, the most current buffer sizing approaches make the assumption that project activity durations are independent. In this paper, the effects of the dependence between activities on project duration performances are analysed, a method for determining buffer sizes with dependence assumption between activities is introduced. Specifically, two definitions reflecting dependence, which are the dependence degree and the dependence factor, are integrated into the formulation of buffer sizing approach. The suggested method is tested and compared with the methods with independence assumption, the root square error method, the adaptive procedure with resource tightness and the adaptive procedure with density. The results indicate that the methods with independence assumption may underestimate the buffer size when at least one of the two definitions is at a high level, while the suggested method can provide better protection in such circumstances.

Keywords:

• project scheduling
• critical chain
• buffer sizing approach
• statistical dependence