Holistic Prescriptive Analytics for Continuous and Constrained Optimization Problems

We present a holistic framework for prescriptive analytics. Given side data x, decisions z, and uncertain quantities y that are functions of x and z, we propose a framework that simultaneously predicts y and prescribes the “should be” optimal decisions $\overline{\mathit{z}}$. The algorithm can accommodate a large number of predictive machine learning models as well as continuous and discrete decisions of high cardinality. It also allows for constraints on these decision variables. We show wide applicability and strong computational performances on synthetic experiments and on two real-world case studies.