distributionally-robust-optimization

Proposes learned predictive ambiguity sets (LPAS) for distributionally robust optimization, where a deep contextual model outputs a nominal scenario distribution, state-dependent Wasserstein radius, and ground metric, trained with decision loss and calibration. Applied to portfolio optimization on S&P 500 data, the method achieves higher returns and Sharpe ratio with reduced conservatism compared to fixed-radius baselines.

This paper proposes a decision-focused generative framework for correlated scenario generation in distributionally robust optimization for grid dispatch, optimizing scenarios based on downstream operational cost rather than forecast accuracy. It reduces operational cost by 0.80–2.02% compared to accuracy-oriented methods across different generative models.