We propose to use Deep Neural Networks to solve data-driven stochastic optimization problems. Given the
historical data of the observed covariate, taken decision, and the realized cost in past periods, we train a
neural network to predict the objective value as a function of the decision and the covariate. Once trained,
for a given covariate, we optimize the neural network over the decision variable using gradient-based methods
because the gradient and the Hessian matrix can be analytically computed. We characterize the performance
of our methodology based on the generalization bound of the neural network. We show strong performance
on two signature problems in operations management, the newsvendor problem and the assortment pricing
problem.
Article
View Using Neural Networks to Guide Data-Driven Operational Decisions