Multiple resources allocation under stochastic conditions on projects with multimodal activities

Abstract

A “multi-modal” activity is one which can be performed in several modes, or levels, of resource(s) allocation. In this paper we assume that the activities are multi-modal, and each activity may require several resources for its execution. We also assume that the resources may be constrained on their availability over time, and that each activity has its own work contents as a continuous random variable for each of its required resources. We are concerned with establishing a suitable strategy that will focus on the universe of activity modes, selecting only those that will be relevant for the most probable overall project cost. In other words, we choose the most probable modes of activity execution among the theoretical continuum of modes; and propagate the consequences of that choice throughout the project. The proposed approach relies on the partitioning of the work content domains into classes. Using these classes, we address different probabilities for each feasible allocation.