Interpretable and Decomposable Multi-Period Convex Risk Measures

Explore a groundbreaking approach to risk assessment in dynamic, uncertain environments through this 41-minute conference talk by Luhao Zhang from the Society for Industrial and Applied Mathematics. Discover how traditional risk measures like Conditional Value-at-Risk fall short due to their lack of dynamic decomposition, and learn about an innovative new class of multi-period convex risk measures that addresses this critical limitation. Understand how this framework evaluates worst-case expectations across all possible stochastic processes while penalizing deviations from nominal processes using both likelihood ratio and outcome-based metrics. Master the key advantage of this approach: its reformulation as a dynamic program that enables more efficient, recursive risk assessment over time. Gain insights into why dynamic decomposition is essential in risk modeling, how this method improves computational efficiency and interpretability, and explore practical applications in finance, operations, and decision-making under uncertainty. Perfect for researchers, practitioners, and students working in quantitative finance, operations research, and stochastic optimization who want to advance their understanding of modern risk assessment techniques.