Data and Risk Analytics for Production Planning

摘要

We examine the classical productional planning model, where a capacity decision that has to be made at the beginning of the planning horizon is the primary means to protect against demand uncertainty. We provide a critique on the model focusing on its profit maximizing objective, its underlying assumptions on demand and related forecasting scheme, and its overall business relevance (or the lack thereof); and we do so in the context of data, risk and analytics. Specifically, we will consider minimizing a shortfall risk relative to a profit target, with a demand model that captures impacts from the financial market and can be learned from data sets that are application specific. With a jointly optimized production and hedging strategy, we show the new model outperforms traditional approaches in risk mitigation as well as in expected profit.