A Deep Reinforcement Learning Heuristic for SAT based on Antagonist Graph Neural Networks.
Thomas FournierArnaud LallouetTélio CropsalGaël GlorianAlexandre PapadopoulosAntoine PetitetGuillaume PerezSuruthy SekarWijnand SuijlenPublished in: ICTAI (2022)
Keyphrases
- neural network
- reinforcement learning
- dynamic programming
- graph theory
- depth first search
- sat solvers
- function approximation
- graph partitioning
- maximum independent set
- graph search
- minimum spanning tree
- graph model
- graph structure
- back propagation
- random walk
- learning algorithm
- genetic algorithm
- artificial neural networks
- pattern recognition
- reinforcement learning algorithms
- combinatorial optimization
- bounded model checking
- state space
- weighted graph
- optimal solution
- tree search
- learning capabilities
- learning classifier systems
- classical planning
- search algorithm
- ai planning
- tabu search
- answer set programming
- simulated annealing
- directed graph
- markov decision processes