First-Order Problem Solving through Neural MCTS based Reinforcement Learning

Xu, Ruiyang; Kadam, Prashank; Lieberherr, Karl

doi:10.48550/arxiv.2101.04167

Cited by 2 publications

(1 citation statement)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…MCTS and Applications MCTS has seen a surge of interest thanks to its great success particularly in solving twoplayer games (Silver et al 2017) (Loth et al 2013) applied MCTS to improve the tree-search heuristics of Constraint Programmin (CP) solvers and showed significant improvements on the depth-first search on certain CP benchmarks. More recent works involve successful application of the "neural" MCTS of AlphaZero to solve a variety of NP-Hard graph problems (Abe et al 2019), as well as solving First-Order Logic descriptions of combinatorial problems (Xu, Kadam, and Lieberherr 2021).…”

Section: Related Workmentioning

confidence: 99%

Finding Backdoors to Integer Programs: A Monte Carlo Tree Search Framework

Khalil

Vaezipoor

Dilkina

2022

AAAI

View full text Add to dashboard Cite

In Mixed Integer Linear Programming (MIP), a (strong) backdoor is a ``small" subset of an instance's integer variables with the following property: in a branch-and-bound procedure, the instance can be solved to global optimality by branching only on the variables in the backdoor. Constructing datasets of pre-computed backdoors for widely used MIP benchmark sets or particular problem families can enable new questions around novel structural properties of a MIP, or explain why a problem that is hard in theory can be solved efficiently in practice. Existing algorithms for finding backdoors rely on sampling candidate variable subsets in various ways, an approach which has demonstrated the existence of backdoors for some instances from MIPLIB2003 and MIPLIB2010. However, these algorithms fall short of consistently succeeding at the task due to an imbalance between exploration and exploitation. We propose BaMCTS, a Monte Carlo Tree Search framework for finding backdoors to MIPs. Extensive algorithmic engineering, hybridization with traditional MIP concepts, and close integration with the CPLEX solver have enabled our method to outperform baselines on MIPLIB2017 instances, finding backdoors more frequently and more efficiently.

show abstract