Stochastic Project Networks

This book presents the state of the art of temporal analysis and cost minimization of projects, as well as project planning under limited resources where the projects are modelled by GERT networks. These projects are characterized by a stochastic evolution structure and possible feedback. The temporal analysis exploits the fact that Markov renewal processes can be associated with so-called EOR networks (special GERT networks all of whose nodes have exclusive-or entrance). For single-machine min-sum and min-max scheduling problems with EOR precedence constraints, two polynomial algorithms are presented. General min-sum problems are solved by dynamic programming. The cost minimization of projects modeled by EOR networks leads to Markov renewal decision processes and can be reduced to a stochastic dynamic programming problem. Basic concepts from the theory of graphs and networks, Markov renewal processes, deterministic scheduling, and dynamic programming are summarized. This makes the book self-contained.