Lookback for Learning to Branch

  title={Lookback for Learning to Branch},
  author={Prateek Gupta and Elias Boutros Khalil and Didier Chet'elat and Maxime Gasse and Yoshua Bengio and Andrea Lodi and M. Pawan Kumar},
The expressive and computationally inexpensive bipartite Graph Neural Networks (GNN) have been shown to be an important component of deep learning based Mixed-Integer Linear Program (MILP) solvers. Recent works have demonstrated the effectiveness of such GNNs in replacing the branching (variable selection) heuristic in branch-and-bound (B&B) solvers. These GNNs are trained, offline and on a collection of MILPs, to imitate a very good but computationally expensive branching heuristic, strong… 

Learning to branch for the crew pairing problem

: Crew pairing problems (CPP) are regularly solved by airlines to produce crew schedules. The goal of CPPs is to find a set of pairings (sequence of flights and rests forming one or more work-days)



Hybrid Models for Learning to Branch

This work addresses the first question in the negative, and addresses the second question by proposing a new hybrid architecture for efficient branching on CPU machines, which combines the expressive power of GNNs with computationally inexpensive multi-layer perceptrons (MLP) for branching.

Parameterizing Branch-and-Bound Search Trees to Learn Branching Policies

A novel imitation learning framework is proposed, and new input features and architectures to represent branching are introduced, and the resulting policies significantly outperform the current state-of-the-art method for "learning to branch" by effectively allowing generalization to generic unseen instances.

Exact Combinatorial Optimization with Graph Convolutional Neural Networks

A new graph convolutional neural network model is proposed for learning branch-and-bound variable selection policies, which leverages the natural variable-constraint bipartite graph representation of mixed-integer linear programs.

Learning to Branch in Mixed Integer Programming

This work proposes a machine learning (ML) framework for variable branching in MIP, and observes the decisions made by Strong Branching, a time-consuming strategy that produces small search trees, collecting features that characterize the candidate branching variables at each node of the tree.

Learning to branch with Tree MDPs

This work revisits the work of Etheve et al. and proposes tree Markov Decision Processes, or tree MDPs, a generalization of temporal M DPs that provides a more suitable framework for learning to branch, and derives a tree policy gradient theorem, which exhibits a better credit assignment compared to its temporal counterpart.

Solving Mixed Integer Programs Using Neural Networks

This paper applies learning to the two key sub-tasks of a MIP solver, generating a high-quality joint variable assignment, and bounding the gap in objective value between that assignment and an optimal one.

Learning to Search in Branch and Bound Algorithms

This work addresses the key challenge of learning an adaptive node searching order for any class of problem solvable by branch-and-bound by applying its algorithm to linear programming based branch- and-bound for solving mixed integer programs (MIP).

Learning to Select Cuts for Efficient Mixed-Integer Programming

Reinforcement Learning for Variable Selection in a Branch and Bound Algorithm

This paper leverage patterns in real-world instances to learn from scratch a new branching strategy optimised for a given problem and compare it with a commercial solver, and proposes FMSTS, a novel Reinforcement Learning approach specifically designed for this task.

Improving Learning to Branch via Reinforcement Learning

This work focuses on learning a variable selection strategy, which is the core of the B&B algorithm, and describes how this strategy is implemented in practice.