Integrated Optimization of Predictive and Prescriptive Tasks
This work is significant for decision-makers who rely on machine learning predictions for subsequent prescriptive actions, by ensuring consistency between predictions and decisions.
This paper addresses the problem of integrating predictive and prescriptive tasks by proposing a new framework that trains predictive algorithm parameters directly within a prescription problem using bilevel optimization. The method aims to prescribe consistent decisions and demonstrates its performance on synthetic data against classical and recent machine learning-based optimization methods.
In traditional machine learning techniques, the degree of closeness between true and predicted values generally measures the quality of predictions. However, these learning algorithms do not consider prescription problems where the predicted values will be used as input to decision problems. In this paper, we efficiently leverage feature variables, and we propose a new framework directly integrating predictive tasks under prescriptive tasks in order to prescribe consistent decisions. We train the parameters of predictive algorithm within a prescription problem via bilevel optimization techniques. We present the structure of our method and demonstrate its performance using synthetic data compared to classical methods like point-estimate-based, stochastic optimization and recently developed machine learning based optimization methods. In addition, we control generalization error using different penalty approaches and optimize the integration over validation data set.