5 Summary, discussion and conclusion
Finding a good estimate of the expected duration (and cost) to develop complex products can be crucial for project managers as they plan for resources and provide clients with estimated completion times. Poor resource planning, staffing decisions, and commitments can result from inaccurate estimates of project duration. Additionally, contractual penalties may arise for noncompliance with promised delivery times. Simulation models to find the expected project duration can be time consuming to design, build, and run. Thus, our focus in this paper is on analytical techniques which can generate the results more promptly.
The cycle elimination (CE) method (Nasr et al., 2016) is an analytical technique which can be used to solve sequential and mixed networks (combination of sequential and parallel activities). The method tackles PD networks having feedback possibilities by transforming the network to a non-cyclic network in order to proceed with traditional project management techniques to find the expected duration. The method uses DSM to represent the network and finds the stage expected durations based on the DSM inputs, where the stages represent the activity's expected duration and their possible rework. Modifications to DSM are necessary for the method capture different complications. Two special cases, coupled activities and allowing parallel rework, were discussed by Nasr et al. (2016) and in this paper we extended the method to account for four additional complications (i) forward probabilities, (ii) dynamic rework probabilities and proportions, (iii) multiple dependency relationships between activities, and (iv) different rework through indirect connections.
