Contextual Scenario Generation for Two-Stage Stochastic Programming
Researchers have developed a new method called Contextual Scenario Generation (CSG) to improve decision-making in two-stage stochastic programs. This technique learns to produce a small set of relevant scenarios based on contextual information, addressing the challenge of needing too many scenarios for accurate approximations. CSG offers two approaches: one focusing on distributional distance and another optimizing decision quality directly. Both methods are broadly applicable and have demonstrated strong empirical performance. AI