Title: Risk-averse decision strategies for influence diagrams using rooted junction trees
Abstract: This paper focuses on a mixed-integer programming formulation for influence diagrams, based on a gradual rooted junction tree representation of the diagram. We show that different risk considerations, including chance constraints and conditional value-at-risk, can be incorporated into the formulation with targeted, appropriate modifications to the diagram structure. The computational performance of the formulation is assessed on two example problems and is found to be highly dependent on the structure of the junction tree.