Scalable Mixed-Integer Optimization with Neural Constraints via Dual Decomposition

Authors

  • Shuli Zeng University of Science and Technology of China
  • Sijia Zhang University of Science and Technology of China
  • Feng Wu University of Science and Technology of China
  • Shaojie Tang State University of New York at Buffalo
  • Xiangyang Li University of Science and Technology of China

DOI:

https://doi.org/10.1609/aaai.v40i17.38454

Abstract

Embedding deep neural networks (NNs) into mixed-integer programs (MIPs) is attractive for decision making with learned constraints, yet state-of-the-art monolithic linearisations blow up in size and quickly become intractable. In this paper, we introduce a novel dual-decomposition framework that relaxes the single coupling equality u=x with an augmented Lagrange multiplier and splits the problem into a vanilla MIP and a constrained NN block. Each part is tackled by the solver that suits it best-branch and cut for the MIP subproblem, first-order optimisation for the NN subproblem, so the model remains modular, the number of integer variables never grows with network depth, and the per-iteration cost scales only linearly with the NN size. On the public SurrogateLIB benchmark, our method proves scalable, modular, and adaptable: it runs 120x faster than an exact Big-M formulation on the largest test case; the NN sub-solver can be swapped from a log-barrier interior step to a projected-gradient routine with no code changes; and swapping the MLP for an LSTM backbone still completes the full optimisation in 47s without any bespoke adaptation.

Published

2026-03-14

How to Cite

Zeng, S., Zhang, S., Wu, F., Tang, S., & Li, X. (2026). Scalable Mixed-Integer Optimization with Neural Constraints via Dual Decomposition. Proceedings of the AAAI Conference on Artificial Intelligence, 40(17), 14388–14396. https://doi.org/10.1609/aaai.v40i17.38454

Issue

Section

AAAI Technical Track on Constraint Satisfaction and Optimization