Fusing Backdoors, Machine Learning, and Optimization for Large-Scale Parametric Mixed-Integer Programs

2026-06-19

AI Summary

This work addresses the high computational cost of repeatedly solving large-scale mixed-integer programming (MIP) problems that share structural similarity but differ in parameters. The authors propose the BIPC framework, which pioneers the integration of backdoor variable identification with machine learning: a supervised learning model predicts values for a critical subset of variables, enabling the construction and solution of a reduced MIP model, followed by a feasibility correction step to recover a high-quality feasible solution. Situated within the “Learning to Optimize” paradigm, this approach significantly reduces solving time in applications such as power systems and transportation—domains requiring rapid responses to parameter perturbations—while incurring only minor losses in solution quality, thereby offering an efficient and practical method for parametric MIP solving.

Abstract

Large-scale optimization problems are often solved repeatedly under similar structural conditions, leading to substantial computational overhead. This occurs in applications such as power systems, transportation, and supply chain networks, where the underlying structure is fixed while parameters frequently vary under perturbations. This paper proposes a Learning to Optimize (LTO) framework that accelerates the solution of large-scale general mixed-integer problems by leveraging the concept of a backdoor, i.e., a subset of variables that drive most of the computational complexity. The proposed BIPC framework consists of three phases. Phase I is an identification procedure that discovers a backdoor for a set of instances in the distribution. Phase II uses supervised learning to develop machine learning models that, given an instance, predict values for bounded-domain backdoor variables and intervals for wide-domain backdoor variables. These predictions define a reduced optimization problem where the predictions constrain the backdoor variables, while the other variables remain free. Phase III optimizes this reduced problem and, if necessary, applies a correction step to restore feasibility or the optimality guarantees. Experiments on real-world, large-scale problems show substantial reductions in solution time with only a limited loss in solution quality. The framework enables organizations to solve large-scale optimization problems efficiently in the presence of frequent perturbations, such as unexpected events, demand fluctuations, or operational changes. Because these changes affect parameters rather than the problem structure, BIPC can quickly provide high-quality, feasible solutions, offering a practical approach to integrating machine learning into existing optimization pipelines.

Problem

• Large-Scale Optimization
• Mixed-Integer Programming
• Parameter Perturbations
• Computational Overhead
• Backdoor Variables

Innovation

• Learning to Optimize
• Backdoor Variables
• Mixed-Integer Programming
• Machine Learning for Optimization
• Parametric Optimization