Abstract:We show that risk-constrained functional optimization problems with general integrable nonconvex instantaneous reward/constraint functions exhibit strong duality, regardless of nonconvexity. We consider risk constraints featuring convex and positively homogeneous risk measures admitting dual representations with bounded risk envelopes, generalizing expectations. Popular risk measures supported within our setting include the conditional value-at-risk (CVaR), the mean-absolute deviation (MAD, including the non-monotone case), certain distributionally robust representations and more generally all real-valued coherent risk measures on the space $L_1$. We highlight the usefulness of our results by further discussing various generalizations of our base model, extensions for risk measures supported on $L_{p>1}$, implications in the context of mean-risk tradeoff models, as well as more specific applications in wireless systems resource allocation, and supervised constrained learning. Our core proof technique appears to be new and relies on risk conjugate duality in tandem with J. J. Uhl's weak extension of A. A. Lyapunov's convexity theorem for vector measures taking values in general infinite-dimensional Banach spaces.

The Neyman–Pearson lemma for convex expectations

A generalized Neyman-Pearson lemma for sublinear expectations

Solving Nonsmooth Nonconvex Compound Stochastic Programs with Applications to Risk Measure Minimization

Uniqueness of Convex-Ranged Probabilities and Applications to Risk Measures and Games

Risk sharing under heterogeneous beliefs without convexity

Strong Duality in Risk-Constrained Nonconvex Functional Programming

Marcinkiewicz's Strong Law of Large Numbers for Nonlinear Expectations

Risk minimization in financial markets modeled by Itô-Lévy processes

On Geometrically Convex Risk Measures

VALUATIONS AND DYNAMIC CONVEX RISK MEASURES

Expectations-based Loss Aversion in Contests

Risk Neutrality Regions

Law-invariant risk measures: Extension properties and qualitative robustness

Risk measures based on weak optimal transport

Conditional Expectations of Correspondences

Probability and convexity concepts are not antagonistic

Dynamic star-shaped risk measures and $g$-expectations

Nonconcave Robust Utility Maximization under Projective Determinacy

Robust convex risk measures

The relations of Choquet Integral and G-Expectation