Mission-driven resource allocation based on subjective input with extra level of uncertainty

This paper presents a solution based on subjective logic to include 'second-order uncertainty' in objective functions for mission-driven resource allocation. When resources have to be optimized, one of the requirements is a probabilistic assessment of future scenarios. However, the end-users who have to convey such future assessment may not be completely confident about it. It can therefore be necessary to specify a confidence level together with this input. With subjective logic it is possible to describe such second-order uncertainty. We used it to describe the uncertainty of assessments (e.g. about future threats) and models (e.g. of sensor systems) within the optimization objective. As a result, the integrity of the optimization improves, because excluding important uncertainties from the optimality definition is not, actually, optimal. In the end, although it is complicated to incorporate different levels of uncertainty, the resulting analysis about mission outcomes is more genuine, it creates transparency towards the end-users, and eventually, results in a resource allocation solution that takes more aspects into account.