Expert Judgments in the Cost-Effectiveness Analysis of Resource Allocations: A Case Study in Military Planning

We consider resource allocation problems in which agents are assigned to tasks with the aim of (i) minimizing the assignment cost and (ii) maximizing the overall value resulting from the completion of tasks. Often, such assignment problems are challenging, because it may not be known to what extent the agents can complete tasks or what the value of either full or partial task completion is. Furthermore, it may be difficult to determine how important the tasks are relative to each other. In this paper, we therefore develop a modeling framework that helps determine for a range of levels of resource expenditure (i) which combinations of agents are cost-effective and (ii) to which tasks these agents should be assigned. The results can be derived by eliciting evaluation judgments from experts. These judgments are analyzed, for example, to determine which combinations of agents outperform others in view of the judgments of all experts, and which ones are cost-ineffective based on the judgments of some or all experts. A computational algorithm is presented and the framework is illustrated by reporting a case study in military planning.