On path floats in a stochastic project network

In a deterministic project network, all the activity times are fixed values and the path float can be found by computing the difference between the project duration and the path length. This is not possible in its stochastic counterpart, where the path lengths and the project duration change constantly since each activity time is a random variable. In the current literature, several methods have been proposed to define the float of a path in this case but all of them are fundamentally flawed. This paper aims to suggest an improved definition of path float in a stochastic project network and show how to compute it through the use of an approximation algorithm for estimating the mean and variance of the project duration under some reasonable assumptions. A numerical example will be provided to demonstrate the superiority of the new approach to the existing ones.