Abstract:The theory of convex risk functions has now been well established as the basis for identifying the families of risk functions that should be used in risk-averse optimization problems. Despite its theoretical appeal, the implementation of a convex risk function remains difficult, because there is little guidance regarding how a convex risk function should be chosen so that it also well represents a decision maker’s subjective risk preference. In this paper, we address this issue through the lens of inverse optimization. Specifically, given solution data from some (forward) risk-averse optimization problem (i.e., a risk minimization problem with known constraints), we develop an inverse optimization framework that generates a risk function that renders the solutions optimal for the forward problem. The framework incorporates the well-known properties of convex risk functions—namely, monotonicity, convexity, translation invariance, and law invariance—as the general information about candidate risk functions, as well as feedback from individuals—which include an initial estimate of the risk function and pairwise comparisons among random losses—as the more specific information. Our framework is particularly novel in that unlike classical inverse optimization, it does not require making any parametric assumption about the risk function (i.e., it is nonparametric). We show how the resulting inverse optimization problems can be reformulated as convex programs and are polynomially solvable if the corresponding forward problems are polynomially solvable. We illustrate the imputed risk functions in a portfolio selection problem and demonstrate their practical value using real-life data. This paper was accepted by Yinyu Ye, optimization.

Strong Duality in Risk-Constrained Nonconvex Functional Programming

Strong Duality Relations in Nonconvex Risk-Constrained Learning

Dual Solutions in Convex Stochastic Optimization

Dual Representation of Unbounded Dynamic Concave Utilities

Strong duality for robust minimax fractional programming problems

Convex duality for stochastic singular control problems

Duality in Regret Measures and Risk Measures

On the dual formulation of regularized linear systems with convex risks

Various duality models associated with some constrained robust nonlinear optimal control problems

A Short and General Duality Proof for Wasserstein Distributionally Robust Optimization

Weak Continuity of Risk Functionals with Applications to Stochastic Programming

M F ] 3 0 A pr 2 01 7 Duality in Regret Measures and Risk Measures

Duality Between Large Deviation Control and Risk-Sensitive Control for Markov Decision Processes.

Lossless Convexification and Duality

A Strong Duality Result for Constrained POMDPs with Multiple Cooperative Agents

On the dual representation of coherent risk measures

General duality and dual attainment for adapted transport

Inverse Optimization of Convex Risk Functions

Unsafe Probabilities and Risk Contours for Stochastic Processes using Convex Optimization

Robust Duality for Fractional Programming Problems with Constraint-Wise Data Uncertainty