On data-driven chance constraint learning for mixed-integer optimization problems

When dealing with real-world optimization problems, decision-makers usually face high levels of uncertainty associated partial information, unknown parameters, or complex relationships between these and the problem decision variables. In this work, we develop a novel Chance Constraint Learning (CCL) methodology focus on mixed-integer linear problems which combines ideas from chance constraint learning literature. constraints set probabilistic confidence level for single to be fulfilled, whereas aims model functional relationship variables through predictive models. One main issues when establishing learned arises need further bounds its response variable: fulfillment is directly related accuracy behaviour. sense, CCL makes use linearizable machine models estimate conditional quantiles variables, providing data-driven solution constraints. An open-access software has been developed used by practitioners. Furthermore, benefits have tested in two case studies, proving how robustness added optimal solutions are