Interval Estimate of Project Time Completion Using Simulation Approach

Abstract

This paper aims to utilize an approach for computing time completion of project network paths with determined activities that can be used to estimate the time completion of a project. Activity durations are considered to be deterministic in critical path method (CPM). Since it is impossible to predict future events in a deterministic manner, project evaluation and review technique (PERT) is introduced which studies project with probabilistic activity durations. In this technique, large variance of path durations can influence the calculation accuracy, in which case the simulation technique is an appropriate method. In this paper, a sample arrow network is chosen, and a time distribution is fitted for each activity with Input Analyzer Software using some historical data. Mean square error, chi-square and Kolmogorov-Smirnov (K-S) tests are used as three measures to evaluate the goodness-of-fit. In the next stage, a computer model is built for the network in Arena 7.0 software. This model is run 120 times and based on these replications, the minimum required sample size for (1-α) Í100% confidence interval is specified. Finally, a confidence interval is computed for duration of each network path and based on this; the time completion of project is estimated.

Keywords

• simulation
• discrete event simulation approach
• project programming