An Approach For Identifying The Most Likely Activity Sequence In A Resource-Constrained Stochastic Network

The biggest concern of a project manager is to identify the proper sequence of activities during project planning and execution so that he or she can appropriately arrange the necessary resources under limited resource availability. The purpose of this article is to develop an approach for identifying the most likely activity sequence in a resource-constrained stochastic network. The approach consists of two stages. The first is a procedure to construct a scenario tree. In the procedure, we integrate a large enough number of sets of feasible activity sequences into a scenario tree. Each set of feasible activity sequences is generated by resource-constrained, deterministic project-scheduling problems with a configuration of randomly generated activity durations. The second stage is a backward-pruning procedure subsequently adopted to eliminate the worse branches from the scenario tree according to a specific objective function. This procedure leads to a reduced scenario tree, on which basis the most likely activity sequence can be easily identified. For demonstration purposes, a simple network example is considered.