Multi-stage resource allocation under uncertainty

In this paper, we discuss a strategic planning problem of allocating resources to groups of tasks organized in successive stages. Each stage is characterized by a set of survival rates whose value is imprecisely known. The goal is to allocate the resources to the tasks (i.e. to form 'teams') by dynamically re-organizing the teams at each stage, while minimizing a cost objective over the whole stage horizon. A modelling framework is proposed, based on linear programming with adjustable variables. The resulting 'uncertain linear program' is subsequently solved using the sampled scenarios randomized technique.