Binary Diversification Characterizes Exact Capacities

Axiomatic Insights into Ambiguity Aversion: Binary Diversification and Exact Capacities In decision sciences, ambiguity refers to uncertainty about event probabilities, and aversion to it is a well-studied phenomenon. In “Binary Diversification Characterizes Exact Capacities,” Hartmann and Kauffeldt consider the Choquet expected utility model, a framework for modeling ambiguity. They provide the first axiomatic characterization of exact capacities, solving an old problem. Their characterizing axiom, binary diversification, captures a specific level of ambiguity aversion—a preference for diversifications that lead to (at most) binary outcomes. They also propose an intuitive hierarchy of ambiguity aversion and illustrate that this hierarchy reflects increasingly strong levels of ambiguity aversion within the more general invariant biseparable model, whereas the hierarchy has limited bite in the Choquet model. They conclude by illustrating the implications for multiobjective shortest-path problems, demonstrating how the results can be applied.