On the Estimate of Mean Activity Durations in PERT Networks

In conventional pert ana lyses, using beta distribution as underlyingdistribut ion for activity durations, the mean and variance of activ ity durations areestim ated using optimistic , pessimistic, and probable estimates give n by an expert.in order to impro ve these estimates, vario us methods have been proposed. in thispaper, we propose a new model, whose estimate of the mean of activity durations ismore acc urate than original pert estimates and its modifications. in addition, theproposed method does not rely on any assumptions on beta parameters.