A bidirectional schedule-improvement procedure for Microsoft Project

Widely-used software packages for project management like Microsoft Project apply simple priority-based procedures for scheduling of resource-constrained projects. Various studies have shown that these procedures perform relatively poor when compared against state-of-the-art procedures. In this paper, we present a scheduling heuristic for Microsoft Project which generates a series of initial schedules by regret-based biased random sampling and improves these schedules by backward-forward passes. In order to consider all types of temporal or calendar constraints that can be modeled in Microsoft Project, both parts of the heuristic employ the built-in schedule-generation scheme of Microsoft Project. Computational results for two real-world construction projects indicate that the presented heuristic improves the resource-allocation capabilities of Microsoft Project considerably.